About

Daniel Edwards

Principal AI Engineering Lead | Context Engineering & Agentic Systems

Director of Engineering focused on AI-first development, system architecture, and modern DevOps practices.

Context EngineeringAgentic SystemsProduction HardeningRegulated Industries

Eighteen-plus years in software, moving from agency web into headless commerce, then into platform leadership and AI engineering. The throughline has always been the same: build systems that hold up when real users hit them, in industries where the eval bar isn't optional.

I specialize in transforming how organizations operate — integrating AI into real-world workflows to increase delivery velocity, improve code quality, and reduce operational friction. My work extends beyond traditional engineering into Context Engineering: designing the systems, data flows, and operational frameworks that enable AI to function reliably at scale. This includes shaping how information is structured, accessed, and governed across an organization so that AI-driven systems can produce consistent, secure, and high-quality outcomes.

I operate at the intersection of AI, platform architecture, and organizational process, helping companies move from traditional development models to scalable, automation-driven ecosystems. Experience includes building secure future-proof architectures, designing enterprise API ecosystems, and implementing AI-assisted engineering pipelines across cloud environments.

Most of my recent work sits in regulated, mission-critical contexts — healthcare (HL7 / FHIR / HIPAA), commerce (PCI), and B2B SaaS (SOC 2). The bar in those rooms isn’t “the demo worked.” It’s:

“This still works at 3am, the audit log is intact, and the model didn’t leak something it shouldn’t have.”

That bar shapes how I think about agent design, prompt-injection defenses, eval-driven CI, and where human oversight belongs in the loop.

I’m particularly focused on making AI practical — embedding it into day-to-day operations and engineering workflows rather than treating it as experimentation. The unglamorous problems are the real work, and the teams that solve them first are the ones whose AI investments actually compound.

Specialty

Context Engineering

Context Engineering is the discipline of designing the data structures, system boundaries, and information flows that let AI agents operate reliably at scale.

Information architecture

Shape how data is structured, retrieved, and ordered so models receive exactly what they need, exactly when they need it.

System boundaries

Define what crosses agent boundaries, what gets summarized vs. preserved verbatim, and how memory tiers (episodic / semantic / working) interact.

Tool surface design

Curate what the model sees, in what order, with what affordances — the tool schema is part of the prompt.

Production governance

Cost ceilings, prompt-injection defenses, eval-driven CI, and the unglamorous work that turns a demo into a system.

Stack

How I actually work

The toolkit I lean on at each phase of the work — and why each one earns its seat. Less about checkboxes, more about how the pieces fit together when an agent has to ship.

Discover

Where the project context actually lives — humans and agents read from the same source.

Confluence

BRDs, FRSs, specs

The single source of truth for project context. When an agent needs to ground itself in why something is being built, this is the page it reads first. If it isn't on a Confluence page, it doesn't exist yet.

Jira

Workflow tracking

Tracking for humans and agents alike. Every ticket carries acceptance criteria; every PR points back to a ticket. The audit trail is built in, not retrofitted later.

Build

The model is the easy part. Routing, IDE integration, and project setup are where the leverage shows up.

Claude Code

Haiku · Sonnet · Opus

No single LLM — each one earns its seat. Haiku for fast scoped edits and grep-style discovery, Sonnet as the daily driver for implementation, Opus for hard architecture and long-context reasoning. Routing the right model to the right task is half the leverage.

  • Windsurf as the IDE

    Primary editor with Claude Code embedded — the agent lives next to the debugger, search, and tests, not in a chat tab in another window.

  • Plugin loadout

    Kept lean: context7, frontend-design, github, skill-creator, feature-dev, claude-md-management, security-guidance. Plugin sprawl is just another kind of context pollution.

  • Project setup

    .claude/settings.local.json scopes permissions and project commands; .mcp.json defines what tools the agent can actually reach. Skip these and you're either babysitting permission prompts all day or giving the agent more access than the project needs.

AI Runtime

How AI shows up inside the product — the SDKs, embeddings, and vector layer the application actually runs on.

Vercel AI SDK / Anthropic SDK

Agent + LLM calls

Vercel AI SDK when streaming UI, tool-use, and React integration matter. Anthropic SDK direct when you don't need the framework abstraction and want the cleanest path to the model. Pick the SDK to fit the use case — same model underneath either way.

OpenAI Embeddings

text-embedding-3-large / small

The default for embedding generation today — quality holds up, latency is fine, pricing is predictable. The model running your product chat doesn't have to be the model running your retrieval index, and shouldn't be.

pgvector

Vector store in Postgres

Vectors live next to the rest of the application data, inside the same Postgres instance and the same security boundary the platform team already audits. The regulated-industries answer: no separate vector DB to threat-model, no extra vendor relationship to negotiate.

Ship

Trunk-based, atomic, traceable. The pipeline enforces the rules so nobody has to remember them.

Azure DevOps

Git + work-item linking

Git management for enterprises that already live in Microsoft's orbit, which is most of them. Repos, pipelines, work-item linking — all in the same place the security and platform teams already approved.

Bicep

Infrastructure-as-Code

IaC for Azure. Terraform is the right answer in a multi-cloud shop; Bicep is the right answer when you're already all-in on Azure and don't want to fight your provider. (Yes — I see what I did there.)

CI Pipelines

Trunk-based deploys

Long-lived feature branches are how merge hell happens. Trunk-based with feature flags and a fast pipeline is how you ship every day without breaking anything — and the pipeline enforces it so nobody has to remember.

Atomic PRs

Traceable to Jira

Each PR does one thing, references one ticket, and stays small enough to actually review. The connective tissue between strategy and code — and the audit trail that makes regulated work possible.

Harden

The production-quality layer. Where "works on my machine" gets replaced by "we have evidence it still works in prod."

Braintrust

Eval-driven CI

Golden datasets, scoring functions, regression catches. Treat AI evals the way you treat unit tests — cheap to run, expensive to skip. Without them, every model change is a guess and every prompt edit is a coin flip.

Langfuse

LLM observability

Traces, costs, prompts, eval feedback — all in one place, OpenTelemetry-native, self-hostable when the data has to stay inside your security boundary. The dashboard you actually look at when production behavior surprises you.

LaunchDarkly

Feature flags

How trunk-based with feature flags actually happens. Wrap every AI feature in a flag — model rollouts, prompt changes, agent toggles — and you can ship hourly without committing to anything irreversible. The kill switch is the feature.

Review & Comm

Humans review judgment. Agents review mechanics. Slack is where the handoff actually happens.

CodeRabbit

Agentic PR review

Catches the easy stuff — style, obvious bugs, missing tests — before a human reviewer ever touches the diff. Humans review judgment; the agent reviews mechanics. The split makes both better.

Slack

Team comms

Where the team actually lives. Threaded context, real handoffs, deployment notifications. Agents post here too — build failures, ticket transitions, deploy status. If it isn't in Slack, it didn't happen.

Harness

The layer above PIV — coordinating the loop. Code-first when engineering owns it; low-code when business teams need to see and edit it.

Archon

Code-first AI harness

Drives execution tracks across the rest of the stack. Where individual agents do focused work inside a PIV loop, Archon coordinates the loop above them — picking tickets, dispatching agents, watching for stalls. The difference between "agent helps with a task" and "agent runs the workflow."

n8n

Low-code companion

When clients need to see and edit the workflow themselves — marketing ops, sales engineers, business teams running their own AI brains — n8n is the answer. Same idea as Archon (orchestrate AI across systems), different audience and a visual canvas non-engineers can actually touch.

Career

What I’ve been building

Fifteen-plus years of progressive technical and leadership work, most-recent first.

AI eraSep 2025 – PresentIrving, TexasHybrid · Full-time

Principal AI Engineering Lead

theAppPlace (TAP) Innovations

  • Lead the Innovative Engineering department, driving AI-first transformation across both internal operations and client-facing solutions; spearhead TAP's evolution from a services-focused org to a hybrid services + product company delivering enterprise applications through white-label and licensing models.
  • Architect and operationalize AI-driven ecosystems across the organization — embedding LLMs, agent-based systems, and automation into engineering, sales, product, project management, and executive decision-making workflows.
  • Pioneer the practice of Context Engineering — designing the data structures, system boundaries, and information flows required for AI systems to operate reliably, securely, and at scale.
  • Define and implement agentic AI workflows, shifting from sequential scripting to reasoning-based execution models that enable adaptive, multi-step problem solving across business functions.
  • Establish governance frameworks for AI adoption — usage controls, auditability, security boundaries, zero-trust data access — ensuring enterprise-grade reliability and compliance.
  • Design and oversee secure RAG architectures, enforcing data access and PII protections while enabling high-quality, context-aware AI outputs.
  • Partner with executive leadership to shape AI strategy, product direction, and company-wide operating models, aligning AI adoption with measurable business outcomes.
Headless commerceNov 2023 – Sep 2025 · 1 yr 11 mosBoston, MassachusettsRemote · Full-time

Senior Technical Architect, BigCommerce Solutions

Verndale

  • Senior technical architect for Verndale's BigCommerce product offering, leading both product design and go-to-market strategy for a new service line.
  • Defined and operationalized delivery workflows across cross-functional teams — engineering, account management, project management, strategy, design, and executive leadership — ensuring alignment from concept through execution.
  • As the AI landscape rapidly evolved, expanded the offering to include AI-driven product R&D and solution development, introducing Windsurf, Contentstack, v0.dev, and Cursor into internal workflows and client-facing capabilities.
  • Led early exploration and adoption of LLM-powered systems and agentic AI workflows, shifting from traditional sequential scripting toward reasoning-based execution models for more adaptive and scalable solutions.
  • Acted as a bridge between technical execution and business strategy, ensuring that emerging AI capabilities were aligned with real-world use cases, delivery models, and client value.
AI eraApr 2023 – Nov 2023 · 8 mosDelray Beach, FloridaRemote · Contract

Engineering Consultant — AI & Platform Systems

PeakActivity

  • Led development teams delivering early-stage AI implementations across both internal and customer-facing systems within eCommerce environments.
  • Directed the design and execution of AI-powered solutions for internal intranet operations and external product experiences, including product selection, recommendation engines, and acquisition flows.
  • Utilized GPT-3.5 and GPT-4 to build and integrate AI-driven capabilities into production workflows — well ahead of widespread enterprise adoption.
  • Balanced hands-on engineering with team leadership, guiding developers through implementation while aligning technical solutions with business objectives.
Headless commerceSep 2021 – Mar 2023 · 1 yr 7 mosDallas, TexasRemote · Full-time

Technical Lead — React Specialist

Rightpoint

  • Led development on AB InBev's BEES global eCommerce platform, supporting large-scale B2B commerce operations across international markets.
  • Served as the technical lead for global language and localization rollouts, owning translation systems and internationalization strategy across regions.
  • Owned and matured DevOps practices, establishing CI/CD standards, deployment strategies, and operational processes for reliable, repeatable releases at scale.
  • Played a key role in adopting BigCommerce as a strategic platform offering; led internal enablement, training and scaling multiple development teams to consistent delivery standards.
  • Architected and guided full-stack solutions across complex commerce ecosystems, integrating backend services, APIs, and frontend applications.
Cloud / DevOpsMay 2019 – Mar 2021 · 1 yr 11 mosProvidence County, Rhode IslandFull-time

Director of Information Technology

Twisted Throttle LLC

  • Responsible for an 8-person IT and Marketing team; managed the eCommerce network of applications and rolled out new sellers channels (Amazon, eBay, Facebook), increasing profits by 20% YoY.
  • Migrated server infrastructure from on-prem to AWS.
  • Managed eCommerce properties on Magento 2, WordPress, and Shopify.
  • Refined the SDLC cycle and implemented a CI/CD strategy.
  • Enhanced ERP integration on NetSuite using a new PostgreSQL database.
Cloud / DevOpsOct 2018 – May 2019 · 8 mosDallas, TexasFull-time

Manager of Application Development

Nobilis Health Corp.

  • Led the design, development, and ongoing evolution of an enterprise healthcare CRM platform supporting patient workflows, insurance communications, and operational approval processes.
  • Architected and maintained integrations aligned with HL7 and FHIR standards, enabling reliable data exchange between internal systems, clinical platforms, and external partners.
  • Directed cross-functional teams building systems that managed insurance verification, claims coordination, and pre-/post-procedure approval workflows.
  • Partnered with business and clinical stakeholders to translate complex regulatory and operational requirements into scalable technical solutions, ensuring HIPAA compliance while maintaining system performance.
  • Reported system health, project progress, and operational risks to executive leadership, supporting strategic planning and resource allocation.

Earlier roles

  • Lead Web Developer

    Globe Runner · Apr 2015 – Jun 2018 · 3 yrs 3 mos

  • Lead Web Developer

    Online Performance Marketing · May 2012 – Feb 2015 · 2 yrs 10 mos

  • Web Developer

    Ilfusion Creative · Jul 2010 – Oct 2011 · 1 yr 4 mos

  • Interactive Production Artist

    e-Rewards · Aug 2008 – Jun 2010 · 1 yr 11 mos

Credentials

Certifications, education & affiliations

The formal track record alongside the artifacts.

Certifications

Cloud & infrastructure

  • Microsoft Azure — AZ-900 (Fundamentals), AZ-204 (Developer Associate)
  • AWS — Cloud Practitioner (CLF-C02), Solutions Architect Associate (SAA)
  • HashiCorp Terraform Associate

Commerce & content platforms

  • BigCommerce BigDev (Developer)
  • BigCommerce Catalyst Developer
  • BigCommerce Implementation / Solutions
  • BigCommerce Headless Specialist
  • Contentstack Certified Developer
  • Contentstack Certified Architect / Champion

AI / LLM

  • Anthropic Skills / course completions

First aid & trauma response

  • American Heart Association (AHA)
  • American Red Cross

Affiliations & community

Industry & AI community

  • Dynamous AI (dynamous.ai) — active community contributor; mentor / advisor; hackathon judge

Freemasonry

  • Master Mason — Ford Lodge No. 270, Grand Lodge of Texas A.F. & A.M.
  • Worshipful Master — current term ending August 2026 (Past Master thereafter)
  • Scottish Rite — 32° Master of the Royal Secret
  • York Rite — Royal Arch · Cryptic Council · Knights Templar
  • Shriners International (AAONMS)
  • Order of the Eastern Star

Recognitions

  • JROTC Medal for Patriotism — National Sojourners

Education

  • B.A., Graphic Design & Multimedia

    Westwood College — Dallas, TX

    2004 – 2008

    3.5 GPA · specialization in web application development, ActionScript, server infrastructure, and ecommerce design.

Endorsements

In their words

Selected recommendations from senior peers, managers, and direct reports.

Daniel is one heck of a web developer lead, manager, mentor, teammate, and a huge asset to a company's success. As my former Manager at Globe Runner and former teammate at Nobilis Health, Daniel has been a tremendous reason I am where I am today. I had the pleasure of having a front row seat in watching Daniel lead organizations' IT departments and how they relied on his expertise, his vision and leadership for success.

Darryl WalkerSenior UX Software Designer · Lockheed Martin

Daniel is an effective, personable and inclusive leader. He inspired his team to do their best work, encouraged collaboration, and to follow process to produce the highest-quality product. In the (too) short time we were on the same team, one of his direct reports was promoted because of the initiative they took, which I take to be a reflection on Daniel's leadership and the environment he created to encourage growth.

Emily AscaniLead Senior Delivery Manager · Genpact

Daniel is an incredibly smart, focused technology leader (and kick ass dev) that transitioned the company from a dying Magento website to a solid WooCommerce solution. He single-handedly cleaned up our complicated deployment process and made it a joy to work with, among a myriad of other friction-easing changes. If you're considering hiring Daniel… HIRE HIM!

Josh SawyerSenior Lead Developer · Wildland Creative

Daniel has such a great grasp of development, tools, and technology. He gets the big picture of how the technology and development/product side could and should work — and he understands his role in achieving business goals. He focuses on the task and has incredible tenacity. I have been consistently impressed with his creativity and his ability to come up with innovative solutions to very difficult problems.

Phil AyresVice President, Marketing · Legent Spine

Daniel is a web dev ninja. If you want a domestic lead dev who has the heart of a teacher, leads by example, and always ends up on the right side of a deadline, Daniel is your man.

Chuck WilliamsPaid Search Consultant · Upwork