The Grid is ready. Illustration unavailable.
HALLUCINATION ALLEY / THE GRID
I fight for the user!
AI product, platform & business builder
I turn complex technology into products, teams, and businesses.
I turn complex technology into products, teams, and businesses.
My work connects product strategy, architecture and enterprise delivery. I have built AI products and solutions architecture teams, then carried that work into customer adoption and commercial partnerships.
Pinpoint Alpha and Keel are independent projects where I explore analytics, AI-assisted software development and governed agent execution.
Building products, teams and enterprise adoption. Start with the latest role, or read the full timeline.
Founding Product Lead, Sovereign AI & Strategic Partnerships
Aug 2025–present
Owned the Sovereign AI product line, helping grow it from prototype to roughly $3M ARR and 40+ global deployments in about 14 months.
Owned the Sovereign AI product line end to end, reporting to the Chief Product Officer. Responsibilities spanned strategy, roadmap, engineering priorities, pricing and packaging, product economics, go-to-market, partnerships, and enterprise delivery.
Led the work with engineers and Solutions Architects to bring Sovereign AI’s self-contained, single-node RKE2 platform into production, with STIG hardening and Ubuntu/RHEL support for enterprise, regulated, and high-security environments.
Set engineering and applied research priorities for the Sovereign AI product line, working across backend engineering, post-training, and inference. Championed FP8/INT4 quantization and support for next-generation GPUs and NVIDIA GB10 edge systems within that product roadmap.
Translated enterprise, regulated, and government customer discovery into Sovereign AI roadmap and go-to-market priorities. Separately led product strategy for Poolside’s $2M In-Q-Tel program spanning four government agencies.
Find me on LinkedIn for professional conversations, or explore my work on GitHub.
Amazon Web Services
NCT3QEEB52QE1QCB
Professional credentials and completed courses. Dates and IDs were supplied by Colin; no renewal is inferred.
Questrom School of Business, Boston University · 2020–2022.
Issued January 2023 · CeDiD: 230A-0542-C8DD.
Independent products and open-source work, from the first idea through working software.
A concise, accessible record of my experience, projects and education.
Product strategy, roadmap ownership, organization design, team scaling, customer discovery, go-to-market planning and strategic partnerships.
Product leadership grounded in architecture and delivery.
Selected public work on enterprise AI and secure deployment.
An efficient path for frontier AI inside your boundary.
Publication · Poolside ·
Authored by Colin Baird for Poolside. An introduction to bringing frontier AI into customer-controlled environments with Dell.
Founding Solutions Architect & On-Premises Lead
Sep 2024–Aug 2025
Poolside’s first Solutions Architect. Built the global team from 0 to 7 and prototyped the on-premises platform.
Joined as employee #47. Established hiring standards, technical playbooks, deployment guides, architecture references and customer-facing runbooks.
Led hands-on architecture, integration and executive stakeholder engagement for the first Fortune 500 customers. Closed a marquee defense prime contract.
Moved the cloud deployment from Terraform and EKS to Docker Compose to make on-premises prototyping possible.
Amazon Web Services
NKK89FY2XEFQQ1WJ
Western Governors University · 2014–2016.
Enterprise AI, foundation models, agent systems, inference, retrieval-augmented generation, Kubernetes (RKE2 and k3s), Terraform, Docker, Python, AWS, Azure and Google Cloud.
Accelerate Government Innovation With Secure, AI-Enabled Models for Software Development (Presented by Poolside)
Talk · NVIDIA GTC DC 2025
A GTC DC 2025 talk on secure AI-enabled software development for government. Session DC51216.
Enterprise Solutions Architect
Oct 2022–Sep 2024
Enterprise architecture and early GenAI: reusable RAG demos, product concepts and strategic cloud migrations.
Authored a December 2022 PR/FAQ for a verticalized Enterprise GenAI Assistant. It was shared with senior AWS AI/ML leadership and contributed to early discussions for what later became Amazon Q.
Led and supported internal and customer-facing technical enablement sessions and demonstrated GenAI prototypes to strategic enterprise customers.
Coursera Course Certificates
JQLWCTCBHJV6
Architecture, technical discovery, proof-of-concept delivery, regulated and air-gapped environments, deployment practices, adoption and partner channels.
Explores why memory may become one of the most important infrastructure layers in AI, especially as models get more capable. I look at where current memory systems fall short and what persistent, domain-specific memory could unlock for software engineering and other long-running work.
Publication · Substack ·
An essay by Colin Baird on Substack.
Senior Solutions Architect
Feb 2022–Aug 2022
Promoted to Senior Solutions Architect in under two years, connecting database architecture with enterprise business value.
Amazon Web Services
NLVDVCT20JQ11P5G
Examines why model intelligence alone may be a weak long-term moat as breakthroughs diffuse and reproduction gets cheaper. I argue that durable AI value is more likely to accrue to the systems around models, including context, governance, orchestration, evaluation, and distribution.
Publication · Substack ·
An essay by Colin Baird on Substack.
Solutions Architect
Jun 2020–Jan 2022
Advised customers on reliable, scalable application and database architectures.
Amazon Web Services
GHD39NYKZEE4143V
Sales Engineer
Apr 2019–Jun 2020
I supported technical sales across a broad infrastructure software portfolio.
Demonstrated, deployed and troubleshot enterprise and commercial solutions spanning DevOps/APM monitoring, security and SIEM, compliance, access rights and patch management.
MongoDB
286-998-837
IT Consultant
Jul 2018–Apr 2019
Enterprise IT consulting across systems, networking and business continuity.
Microsoft
I010-4104
IT Specialist II
Oct 2015–Jul 2018
Hands-on infrastructure consulting and technical support for small and midsize businesses.
MongoDB
989-793-551
IT Support Engineer
Mar 2015–Oct 2015
Enterprise IT support across campus, research centers and 254 county offices.
Amazon Web Services
EYBHE9DB12V118S1
SolarWinds
2100
An AI-native market research SaaS that turns institutional flow and market structure into a clear, personalized daily read.
I owned product strategy, architecture and delivery across market-data pipelines, analytics, the web experience and subscriptions. Took the product from concept to production in three months.
Python computes and scores market signals before the model sees them. Options flow, dark-pool activity, open interest, volatility and dealer positioning feed a deterministic evidence layer; AI explains that evidence in plain English.
The engineering work includes provenance, freshness checks and fail-closed quality gates. Regression tests cover stale data, mismatched contracts and incomplete evidence so unsupported claims do not quietly reach the user.
Next.js on Vercel serves the product; Supabase provides the data and authentication layer; Railway runs Python pipelines. Massive supplies market data, Anthropic powers the explanation layer, Stripe manages subscriptions, and Resend delivers the briefs.
A personalized evening brief connects to supporting dashboards, institutional price levels, sector rotation and position tracking. Progressive disclosure keeps the first read concise while preserving the evidence underneath.
Built for research and education. Signal quality and evidence matter more than promises of trading returns.
XnIhwu
An open-source coding-agent harness built around a hard question: how do you give an AI autonomy without giving it authority over its own rules?
I defined the architecture, threat model, contracts and evidence gates, using AI extensively for implementation, testing and documentation.
A separate Warden process owns policy decisions, sandbox execution and audit writes. The agent requests actions; it cannot approve itself or rewrite its hash-pinned policy.
Side effects require a durable intent record before execution. Allowed and denied actions produce tamper-evident evidence; failed enforcement checks stop execution.
Shell commands, typed file tools and reviewed local MCP tools share the same enforcement path. Publishing code and opening pull requests require separate, narrowly scoped approvals.
Pre-alpha, solo-maintained and AI-assisted. An unaudited learning project and executable security reference, not a production-ready security guarantee. Its boundary assumes a trusted kernel and OS account.
Coursera Course Certificates
C2TL78JWRCRF
Co-developed at AWS in February 2023: a reusable, open-source chatbot that connects enterprise documents to conversational AI, from early GenAI experimentation to a customizable deployment.
I co-developed the original work at AWS in February 2023, as teams were beginning to explore how retrieval-augmented generation could make language models useful with their own documents.
The linked Bedrock implementation combines Amazon Bedrock for inference, Amazon Kendra for document retrieval and Amazon Lex for conversation. Python, LangChain and AWS Lambda connect the services; CloudFormation makes the environment repeatable.
Lex handles known intents directly and routes open-ended questions through retrieval and generation. Conversation history supports follow-up questions, while Lex integration with Amazon Connect brings the approach into support-center workflows.
Teams can change the source documents and swap models instead of starting from scratch. The value was making an emerging AI pattern concrete: deploy it, inspect the architecture, and adapt it to a real customer problem.
Coursera Course Certificates
NBYCAPZW2QBC
LGGHGBALWNCP
Coursera Course Certificates
EMKEMFC7V2PJ
Coursera Course Certificates
8B25P6JP3ED9
CompTIA
J1Q7METGP3Q1K994
I turn complex technology into products, teams, and businesses.
My work connects product strategy, architecture and enterprise delivery. I have built AI products and solutions architecture teams, then carried that work into customer adoption and commercial partnerships.
Pinpoint Alpha and Keel are independent projects where I explore analytics, AI-assisted software development and governed agent execution.
Founding Product Lead, Sovereign AI & Strategic Partnerships
Aug 2025–present
Owned the Sovereign AI product line, helping grow it from prototype to roughly $3M ARR and 40+ global deployments in about 14 months.
Owned the Sovereign AI product line end to end, reporting to the Chief Product Officer. Responsibilities spanned strategy, roadmap, engineering priorities, pricing and packaging, product economics, go-to-market, partnerships, and enterprise delivery.
Led the work with engineers and Solutions Architects to bring Sovereign AI’s self-contained, single-node RKE2 platform into production, with STIG hardening and Ubuntu/RHEL support for enterprise, regulated, and high-security environments.
Set engineering and applied research priorities for the Sovereign AI product line, working across backend engineering, post-training, and inference. Championed FP8/INT4 quantization and support for next-generation GPUs and NVIDIA GB10 edge systems within that product roadmap.
Translated enterprise, regulated, and government customer discovery into Sovereign AI roadmap and go-to-market priorities. Separately led product strategy for Poolside’s $2M In-Q-Tel program spanning four government agencies.
Amazon Web Services
NCT3QEEB52QE1QCB
Questrom School of Business, Boston University · 2020–2022.
Issued January 2023 · CeDiD: 230A-0542-C8DD.
Product strategy, roadmap ownership, organization design, team scaling, customer discovery, go-to-market planning and strategic partnerships.
An efficient path for frontier AI inside your boundary.
Publication · Poolside ·
Authored by Colin Baird for Poolside. An introduction to bringing frontier AI into customer-controlled environments with Dell.
Founding Solutions Architect & On-Premises Lead
Sep 2024–Aug 2025
Poolside’s first Solutions Architect. Built the global team from 0 to 7 and prototyped the on-premises platform.
Joined as employee #47. Established hiring standards, technical playbooks, deployment guides, architecture references and customer-facing runbooks.
Led hands-on architecture, integration and executive stakeholder engagement for the first Fortune 500 customers. Closed a marquee defense prime contract.
Moved the cloud deployment from Terraform and EKS to Docker Compose to make on-premises prototyping possible.
Amazon Web Services
NKK89FY2XEFQQ1WJ
Western Governors University · 2014–2016.
Enterprise AI, foundation models, agent systems, inference, retrieval-augmented generation, Kubernetes (RKE2 and k3s), Terraform, Docker, Python, AWS, Azure and Google Cloud.
Accelerate Government Innovation With Secure, AI-Enabled Models for Software Development (Presented by Poolside)
Talk · NVIDIA GTC DC 2025
A GTC DC 2025 talk on secure AI-enabled software development for government. Session DC51216.
Enterprise Solutions Architect
Oct 2022–Sep 2024
Enterprise architecture and early GenAI: reusable RAG demos, product concepts and strategic cloud migrations.
Authored a December 2022 PR/FAQ for a verticalized Enterprise GenAI Assistant. It was shared with senior AWS AI/ML leadership and contributed to early discussions for what later became Amazon Q.
Led and supported internal and customer-facing technical enablement sessions and demonstrated GenAI prototypes to strategic enterprise customers.
Coursera Course Certificates
JQLWCTCBHJV6
Architecture, technical discovery, proof-of-concept delivery, regulated and air-gapped environments, deployment practices, adoption and partner channels.
Explores why memory may become one of the most important infrastructure layers in AI, especially as models get more capable. I look at where current memory systems fall short and what persistent, domain-specific memory could unlock for software engineering and other long-running work.
Publication · Substack ·
An essay by Colin Baird on Substack.
Senior Solutions Architect
Feb 2022–Aug 2022
Promoted to Senior Solutions Architect in under two years, connecting database architecture with enterprise business value.
Amazon Web Services
NLVDVCT20JQ11P5G
Examines why model intelligence alone may be a weak long-term moat as breakthroughs diffuse and reproduction gets cheaper. I argue that durable AI value is more likely to accrue to the systems around models, including context, governance, orchestration, evaluation, and distribution.
Publication · Substack ·
An essay by Colin Baird on Substack.
Solutions Architect
Jun 2020–Jan 2022
Advised customers on reliable, scalable application and database architectures.
Amazon Web Services
GHD39NYKZEE4143V
Sales Engineer
Apr 2019–Jun 2020
I supported technical sales across a broad infrastructure software portfolio.
Demonstrated, deployed and troubleshot enterprise and commercial solutions spanning DevOps/APM monitoring, security and SIEM, compliance, access rights and patch management.
MongoDB
286-998-837
IT Consultant
Jul 2018–Apr 2019
Enterprise IT consulting across systems, networking and business continuity.
Microsoft
I010-4104
IT Specialist II
Oct 2015–Jul 2018
Hands-on infrastructure consulting and technical support for small and midsize businesses.
MongoDB
989-793-551
IT Support Engineer
Mar 2015–Oct 2015
Enterprise IT support across campus, research centers and 254 county offices.
Amazon Web Services
EYBHE9DB12V118S1
SolarWinds
2100
An AI-native market research SaaS that turns institutional flow and market structure into a clear, personalized daily read.
I owned product strategy, architecture and delivery across market-data pipelines, analytics, the web experience and subscriptions. Took the product from concept to production in three months.
Python computes and scores market signals before the model sees them. Options flow, dark-pool activity, open interest, volatility and dealer positioning feed a deterministic evidence layer; AI explains that evidence in plain English.
The engineering work includes provenance, freshness checks and fail-closed quality gates. Regression tests cover stale data, mismatched contracts and incomplete evidence so unsupported claims do not quietly reach the user.
Next.js on Vercel serves the product; Supabase provides the data and authentication layer; Railway runs Python pipelines. Massive supplies market data, Anthropic powers the explanation layer, Stripe manages subscriptions, and Resend delivers the briefs.
A personalized evening brief connects to supporting dashboards, institutional price levels, sector rotation and position tracking. Progressive disclosure keeps the first read concise while preserving the evidence underneath.
Built for research and education. Signal quality and evidence matter more than promises of trading returns.
XnIhwu
An open-source coding-agent harness built around a hard question: how do you give an AI autonomy without giving it authority over its own rules?
I defined the architecture, threat model, contracts and evidence gates, using AI extensively for implementation, testing and documentation.
A separate Warden process owns policy decisions, sandbox execution and audit writes. The agent requests actions; it cannot approve itself or rewrite its hash-pinned policy.
Side effects require a durable intent record before execution. Allowed and denied actions produce tamper-evident evidence; failed enforcement checks stop execution.
Shell commands, typed file tools and reviewed local MCP tools share the same enforcement path. Publishing code and opening pull requests require separate, narrowly scoped approvals.
Pre-alpha, solo-maintained and AI-assisted. An unaudited learning project and executable security reference, not a production-ready security guarantee. Its boundary assumes a trusted kernel and OS account.
Coursera Course Certificates
C2TL78JWRCRF
Co-developed at AWS in February 2023: a reusable, open-source chatbot that connects enterprise documents to conversational AI, from early GenAI experimentation to a customizable deployment.
I co-developed the original work at AWS in February 2023, as teams were beginning to explore how retrieval-augmented generation could make language models useful with their own documents.
The linked Bedrock implementation combines Amazon Bedrock for inference, Amazon Kendra for document retrieval and Amazon Lex for conversation. Python, LangChain and AWS Lambda connect the services; CloudFormation makes the environment repeatable.
Lex handles known intents directly and routes open-ended questions through retrieval and generation. Conversation history supports follow-up questions, while Lex integration with Amazon Connect brings the approach into support-center workflows.
Teams can change the source documents and swap models instead of starting from scratch. The value was making an emerging AI pattern concrete: deploy it, inspect the architecture, and adapt it to a real customer problem.
Coursera Course Certificates
NBYCAPZW2QBC
LGGHGBALWNCP
Coursera Course Certificates
EMKEMFC7V2PJ
Coursera Course Certificates
8B25P6JP3ED9
CompTIA
J1Q7METGP3Q1K994
CHAPTER 01
Founding Product Lead, Sovereign AI & Strategic Partnerships
Aug 2025–present
Owned the Sovereign AI product line, helping grow it from prototype to roughly $3M ARR and 40+ global deployments in about 14 months.
Owned the Sovereign AI product line end to end, reporting to the Chief Product Officer. Responsibilities spanned strategy, roadmap, engineering priorities, pricing and packaging, product economics, go-to-market, partnerships, and enterprise delivery.
Led the work with engineers and Solutions Architects to bring Sovereign AI’s self-contained, single-node RKE2 platform into production, with STIG hardening and Ubuntu/RHEL support for enterprise, regulated, and high-security environments.
Set engineering and applied research priorities for the Sovereign AI product line, working across backend engineering, post-training, and inference. Championed FP8/INT4 quantization and support for next-generation GPUs and NVIDIA GB10 edge systems within that product roadmap.
Translated enterprise, regulated, and government customer discovery into Sovereign AI roadmap and go-to-market priorities. Separately led product strategy for Poolside’s $2M In-Q-Tel program spanning four government agencies.
CHAPTER 02
Founding Solutions Architect & On-Premises Lead
Sep 2024–Aug 2025
Poolside’s first Solutions Architect. Built the global team from 0 to 7 and prototyped the on-premises platform.
Joined as employee #47. Established hiring standards, technical playbooks, deployment guides, architecture references and customer-facing runbooks.
Led hands-on architecture, integration and executive stakeholder engagement for the first Fortune 500 customers. Closed a marquee defense prime contract.
Moved the cloud deployment from Terraform and EKS to Docker Compose to make on-premises prototyping possible.
CHAPTER 03
Enterprise Solutions Architect
Oct 2022–Sep 2024
Enterprise architecture and early GenAI: reusable RAG demos, product concepts and strategic cloud migrations.
Authored a December 2022 PR/FAQ for a verticalized Enterprise GenAI Assistant. It was shared with senior AWS AI/ML leadership and contributed to early discussions for what later became Amazon Q.
Led and supported internal and customer-facing technical enablement sessions and demonstrated GenAI prototypes to strategic enterprise customers.
CHAPTER 04
Senior Solutions Architect
Feb 2022–Aug 2022
Promoted to Senior Solutions Architect in under two years, connecting database architecture with enterprise business value.
CHAPTER 05
Solutions Architect
Jun 2020–Jan 2022
Advised customers on reliable, scalable application and database architectures.
CHAPTER 06
Sales Engineer
Apr 2019–Jun 2020
I supported technical sales across a broad infrastructure software portfolio.
Demonstrated, deployed and troubleshot enterprise and commercial solutions spanning DevOps/APM monitoring, security and SIEM, compliance, access rights and patch management.
CHAPTER 07
IT Consultant
Jul 2018–Apr 2019
Enterprise IT consulting across systems, networking and business continuity.
CHAPTER 08
IT Specialist II
Oct 2015–Jul 2018
Hands-on infrastructure consulting and technical support for small and midsize businesses.
CHAPTER 09
IT Support Engineer
Mar 2015–Oct 2015
Enterprise IT support across campus, research centers and 254 county offices.
← → to browse while this archive has focus. Home / End jump to first / last. Tab moves to the next control.
CHAPTER 01
Founding Product Lead, Sovereign AI & Strategic Partnerships
Aug 2025–present
Owned the Sovereign AI product line, helping grow it from prototype to roughly $3M ARR and 40+ global deployments in about 14 months.
Owned the Sovereign AI product line end to end, reporting to the Chief Product Officer. Responsibilities spanned strategy, roadmap, engineering priorities, pricing and packaging, product economics, go-to-market, partnerships, and enterprise delivery.
Led the work with engineers and Solutions Architects to bring Sovereign AI’s self-contained, single-node RKE2 platform into production, with STIG hardening and Ubuntu/RHEL support for enterprise, regulated, and high-security environments.
Set engineering and applied research priorities for the Sovereign AI product line, working across backend engineering, post-training, and inference. Championed FP8/INT4 quantization and support for next-generation GPUs and NVIDIA GB10 edge systems within that product roadmap.
Translated enterprise, regulated, and government customer discovery into Sovereign AI roadmap and go-to-market priorities. Separately led product strategy for Poolside’s $2M In-Q-Tel program spanning four government agencies.
CHAPTER 02
Founding Solutions Architect & On-Premises Lead
Sep 2024–Aug 2025
Poolside’s first Solutions Architect. Built the global team from 0 to 7 and prototyped the on-premises platform.
Joined as employee #47. Established hiring standards, technical playbooks, deployment guides, architecture references and customer-facing runbooks.
Led hands-on architecture, integration and executive stakeholder engagement for the first Fortune 500 customers. Closed a marquee defense prime contract.
Moved the cloud deployment from Terraform and EKS to Docker Compose to make on-premises prototyping possible.
CHAPTER 03
Enterprise Solutions Architect
Oct 2022–Sep 2024
Enterprise architecture and early GenAI: reusable RAG demos, product concepts and strategic cloud migrations.
Authored a December 2022 PR/FAQ for a verticalized Enterprise GenAI Assistant. It was shared with senior AWS AI/ML leadership and contributed to early discussions for what later became Amazon Q.
Led and supported internal and customer-facing technical enablement sessions and demonstrated GenAI prototypes to strategic enterprise customers.
CHAPTER 04
Senior Solutions Architect
Feb 2022–Aug 2022
Promoted to Senior Solutions Architect in under two years, connecting database architecture with enterprise business value.
CHAPTER 05
Solutions Architect
Jun 2020–Jan 2022
Advised customers on reliable, scalable application and database architectures.
CHAPTER 06
Sales Engineer
Apr 2019–Jun 2020
I supported technical sales across a broad infrastructure software portfolio.
Demonstrated, deployed and troubleshot enterprise and commercial solutions spanning DevOps/APM monitoring, security and SIEM, compliance, access rights and patch management.
CHAPTER 07
IT Consultant
Jul 2018–Apr 2019
Enterprise IT consulting across systems, networking and business continuity.
CHAPTER 08
IT Specialist II
Oct 2015–Jul 2018
Hands-on infrastructure consulting and technical support for small and midsize businesses.
CHAPTER 09
IT Support Engineer
Mar 2015–Oct 2015
Enterprise IT support across campus, research centers and 254 county offices.
← → to browse while this archive has focus. Home / End jump to first / last. Tab moves to the next control.
A place to pause. A story to explore.
← → Browse · ↑ ↓ Read · Escape Return to city
A place to pause. A story to explore.
← → Browse · ↑ ↓ Read · Escape Return to city
A place to pause. A story to explore.
← → Browse · ↑ ↓ Read · Escape Return to city
A place to pause. A story to explore.
PROJECT
An AI-native market research SaaS that turns institutional flow and market structure into a clear, personalized daily read.
Read Pinpoint Alpha ↗PROJECT
An open-source coding-agent harness built around a hard question: how do you give an AI autonomy without giving it authority over its own rules?
Read Keel ↗PROJECT
Co-developed at AWS in February 2023: a reusable, open-source chatbot that connects enterprise documents to conversational AI, from early GenAI experimentation to a customizable deployment.
Read AWS RAG Chatbot ↗WRITING
An efficient path for frontier AI inside your boundary.
Read Poolside on Dell ↗WRITING
Accelerate Government Innovation With Secure, AI-Enabled Models for Software Development (Presented by Poolside)
Read NVIDIA GTC DC 2025 ↗WRITING
Explores why memory may become one of the most important infrastructure layers in AI, especially as models get more capable. I look at where current memory systems fall short and what persistent, domain-specific memory could unlock for software engineering and other long-running work.
Read The Six-Hour Mind: Why Intelligence Can’t Compound Without Memory ↗WRITING
Examines why model intelligence alone may be a weak long-term moat as breakthroughs diffuse and reproduction gets cheaper. I argue that durable AI value is more likely to accrue to the systems around models, including context, governance, orchestration, evaluation, and distribution.
Read AI Moats Through the Lens of Multiple Discovery ↗SKILL
Product strategy, roadmap ownership, organization design, team scaling, customer discovery, go-to-market planning and strategic partnerships.
Read Product & leadership ↗SKILL
Enterprise AI, foundation models, agent systems, inference, retrieval-augmented generation, Kubernetes (RKE2 and k3s), Terraform, Docker, Python, AWS, Azure and Google Cloud.
Read AI & platforms ↗SKILL
Architecture, technical discovery, proof-of-concept delivery, regulated and air-gapped environments, deployment practices, adoption and partner channels.
Read Enterprise delivery ↗Status as of 2026-09-21. “No expiry supplied” means no expiry date was provided; it does not assert an active certification. Credential IDs are available below.
Amazon Web Services · Issued Nov 2023 · Expires Nov 2026
NCT3QEEB52QE1QCB
Amazon Web Services · Issued Nov 2023 · Expires Nov 2026
NKK89FY2XEFQQ1WJ
Coursera Course Certificates · Issued Sep 2023
JQLWCTCBHJV6
MongoDB · Issued Apr 2022
286-998-837
Microsoft · Issued Oct 2021
I010-4104
MongoDB · Issued Jan 2021
989-793-551
Coursera Course Certificates · Issued Jan 2020
C2TL78JWRCRF
Coursera Course Certificates · Issued Dec 2019
NBYCAPZW2QBC
Google · Issued Jul 2019
LGGHGBALWNCP
Coursera Course Certificates · Issued May 2019
EMKEMFC7V2PJ
Coursera Course Certificates · Issued Apr 2019
8B25P6JP3ED9
CompTIA · Issued Jun 2016
J1Q7METGP3Q1K994
Amazon Web Services · Issued Jun 2023 · Expired Jun 2026
NLVDVCT20JQ11P5G
Amazon Web Services · Issued Mar 2023 · Expired Mar 2026
GHD39NYKZEE4143V
Amazon Web Services · Issued Oct 2020 · Expired Oct 2023
EYBHE9DB12V118S1
SolarWinds · Issued Mar 2020 · Expired Mar 2023
2100
Google · Issued Jan 2020 · Expired Jan 2022
XnIhwu
Content reviewed
Download résumé PDFI turn complex technology into products, teams, and businesses.
Founding Product Lead, Sovereign AI & Strategic Partnerships
Aug 2025–present
Owned the Sovereign AI product line, helping grow it from prototype to roughly $3M ARR and 40+ global deployments in about 14 months.
Owned the Sovereign AI product line end to end, reporting to the Chief Product Officer. Responsibilities spanned strategy, roadmap, engineering priorities, pricing and packaging, product economics, go-to-market, partnerships, and enterprise delivery.
Led the work with engineers and Solutions Architects to bring Sovereign AI’s self-contained, single-node RKE2 platform into production, with STIG hardening and Ubuntu/RHEL support for enterprise, regulated, and high-security environments.
Set engineering and applied research priorities for the Sovereign AI product line, working across backend engineering, post-training, and inference. Championed FP8/INT4 quantization and support for next-generation GPUs and NVIDIA GB10 edge systems within that product roadmap.
Translated enterprise, regulated, and government customer discovery into Sovereign AI roadmap and go-to-market priorities. Separately led product strategy for Poolside’s $2M In-Q-Tel program spanning four government agencies.
Founding Solutions Architect & On-Premises Lead
Sep 2024–Aug 2025
Poolside’s first Solutions Architect. Built the global team from 0 to 7 and prototyped the on-premises platform.
Joined as employee #47. Established hiring standards, technical playbooks, deployment guides, architecture references and customer-facing runbooks.
Led hands-on architecture, integration and executive stakeholder engagement for the first Fortune 500 customers. Closed a marquee defense prime contract.
Moved the cloud deployment from Terraform and EKS to Docker Compose to make on-premises prototyping possible.
Enterprise Solutions Architect
Oct 2022–Sep 2024
Enterprise architecture and early GenAI: reusable RAG demos, product concepts and strategic cloud migrations.
Authored a December 2022 PR/FAQ for a verticalized Enterprise GenAI Assistant. It was shared with senior AWS AI/ML leadership and contributed to early discussions for what later became Amazon Q.
Led and supported internal and customer-facing technical enablement sessions and demonstrated GenAI prototypes to strategic enterprise customers.
Senior Solutions Architect
Feb 2022–Aug 2022
Promoted to Senior Solutions Architect in under two years, connecting database architecture with enterprise business value.
Solutions Architect
Jun 2020–Jan 2022
Advised customers on reliable, scalable application and database architectures.
Sales Engineer
Apr 2019–Jun 2020
I supported technical sales across a broad infrastructure software portfolio.
Demonstrated, deployed and troubleshot enterprise and commercial solutions spanning DevOps/APM monitoring, security and SIEM, compliance, access rights and patch management.
IT Consultant
Jul 2018–Apr 2019
Enterprise IT consulting across systems, networking and business continuity.
IT Specialist II
Oct 2015–Jul 2018
Hands-on infrastructure consulting and technical support for small and midsize businesses.
IT Support Engineer
Mar 2015–Oct 2015
Enterprise IT support across campus, research centers and 254 county offices.
An AI-native market research SaaS that turns institutional flow and market structure into a clear, personalized daily read.
I owned product strategy, architecture and delivery across market-data pipelines, analytics, the web experience and subscriptions. Took the product from concept to production in three months.
Python computes and scores market signals before the model sees them. Options flow, dark-pool activity, open interest, volatility and dealer positioning feed a deterministic evidence layer; AI explains that evidence in plain English.
The engineering work includes provenance, freshness checks and fail-closed quality gates. Regression tests cover stale data, mismatched contracts and incomplete evidence so unsupported claims do not quietly reach the user.
Next.js on Vercel serves the product; Supabase provides the data and authentication layer; Railway runs Python pipelines. Massive supplies market data, Anthropic powers the explanation layer, Stripe manages subscriptions, and Resend delivers the briefs.
A personalized evening brief connects to supporting dashboards, institutional price levels, sector rotation and position tracking. Progressive disclosure keeps the first read concise while preserving the evidence underneath.
Built for research and education. Signal quality and evidence matter more than promises of trading returns.
An open-source coding-agent harness built around a hard question: how do you give an AI autonomy without giving it authority over its own rules?
I defined the architecture, threat model, contracts and evidence gates, using AI extensively for implementation, testing and documentation.
A separate Warden process owns policy decisions, sandbox execution and audit writes. The agent requests actions; it cannot approve itself or rewrite its hash-pinned policy.
Side effects require a durable intent record before execution. Allowed and denied actions produce tamper-evident evidence; failed enforcement checks stop execution.
Shell commands, typed file tools and reviewed local MCP tools share the same enforcement path. Publishing code and opening pull requests require separate, narrowly scoped approvals.
Pre-alpha, solo-maintained and AI-assisted. An unaudited learning project and executable security reference, not a production-ready security guarantee. Its boundary assumes a trusted kernel and OS account.
Co-developed at AWS in February 2023: a reusable, open-source chatbot that connects enterprise documents to conversational AI, from early GenAI experimentation to a customizable deployment.
I co-developed the original work at AWS in February 2023, as teams were beginning to explore how retrieval-augmented generation could make language models useful with their own documents.
The linked Bedrock implementation combines Amazon Bedrock for inference, Amazon Kendra for document retrieval and Amazon Lex for conversation. Python, LangChain and AWS Lambda connect the services; CloudFormation makes the environment repeatable.
Lex handles known intents directly and routes open-ended questions through retrieval and generation. Conversation history supports follow-up questions, while Lex integration with Amazon Connect brings the approach into support-center workflows.
Teams can change the source documents and swap models instead of starting from scratch. The value was making an emerging AI pattern concrete: deploy it, inspect the architecture, and adapt it to a real customer problem.
Questrom School of Business, Boston University · 2020–2022.
Western Governors University · 2014–2016.
Product strategy, roadmap ownership, organization design, team scaling, customer discovery, go-to-market planning and strategic partnerships.
Enterprise AI, foundation models, agent systems, inference, retrieval-augmented generation, Kubernetes (RKE2 and k3s), Terraform, Docker, Python, AWS, Azure and Google Cloud.
Architecture, technical discovery, proof-of-concept delivery, regulated and air-gapped environments, deployment practices, adoption and partner channels.
An efficient path for frontier AI inside your boundary.
Publication · Poolside ·
Authored by Colin Baird for Poolside. An introduction to bringing frontier AI into customer-controlled environments with Dell.
Accelerate Government Innovation With Secure, AI-Enabled Models for Software Development (Presented by Poolside)
Talk · NVIDIA GTC DC 2025
A GTC DC 2025 talk on secure AI-enabled software development for government. Session DC51216.
Explores why memory may become one of the most important infrastructure layers in AI, especially as models get more capable. I look at where current memory systems fall short and what persistent, domain-specific memory could unlock for software engineering and other long-running work.
Publication · Substack ·
An essay by Colin Baird on Substack.
Examines why model intelligence alone may be a weak long-term moat as breakthroughs diffuse and reproduction gets cheaper. I argue that durable AI value is more likely to accrue to the systems around models, including context, governance, orchestration, evaluation, and distribution.
Publication · Substack ·
An essay by Colin Baird on Substack.
Find me on LinkedIn for professional conversations, or explore my work on GitHub.
I turn complex technology into products, teams, and businesses.
My work connects product strategy, architecture and enterprise delivery. I have built AI products and solutions architecture teams, then carried that work into customer adoption and commercial partnerships.
Pinpoint Alpha and Keel are independent projects where I explore analytics, AI-assisted software development and governed agent execution.
Fictional plaza guide
Open Career in the navigation for the full timeline. In the city, Career Avenue leads to the Career History building. Its archive has one chapter per role, with arrow keys to browse.
Explore this topic ↗For a quick overview, use Career and Projects in the portfolio navigation. Keel and Pinpoint Alpha also have entrances at Technology Park. The city is optional; every topic has a regular page.
Explore this topic ↗These are three locally scripted answers from a fictional guide. There is no AI call, chat memory, or account. Good directions don’t need a data center.
Explore this topic ↗Three authored answers. No AI, account, or conversation storage.
1 credit per item. Everything is served lukewarm on the first request.
Three orders per visit. The barista rate-limits.
Two things on the board. Both take a moment.
It is always DNS.
Progress lives in this browser only and is never sent anywhere. Reset to plaza keeps it.
GIT A PAINT JOB LATENCY LANE
← → Browse · Enter Fit · Esc Drive out
Appearance only. Same ride, same handling.
Saved in this browser. Reset to plaza keeps your builds and credits.
Taking your seat…
Watch on YouTube ↗A tribute to the communities behind Reddit’s shared pixel-art canvas.
Loading the canvas…
The portfolio is always one click away.
The city is silent. Cinema sound uses the YouTube player controls. Hard vehicle-to-vehicle impacts add brief light smoke. Seven separate hard impacts add lasting cosmetic wear. Reset to plaza clears both.
