Staff Engineer · AI Platform & Developer Enablement

Derrick Bryant

I bring agentic development practices into engineering organizations.

For 18 years I've built the platforms, pipelines, and cross-team practices that let developers focus on creating value instead of fighting their tooling — now with AI agents at the center. I design for failure, reduce cognitive load, and build systems people actually trust.

About

An engineer who makes new technology stick

I'm a Staff-level engineer who comes into an organization and raises how fast — and how safely — it ships. Most of my career has lived at the intersection of development and infrastructure: I architect the platforms and practices that let teams move quickly without breaking trust. At CareJourney I took deployments from monthly, high-risk events to multiple routine releases a day, ran a company-wide shift-left initiative, and built AI-powered internal tools — all in a HIPAA-compliant environment.

The hard part of adopting new technology — especially agentic AI — isn't the demo. It's making it reliable, governed, and actually adopted across teams. That's the exact motion I've run for years: stand up the tooling, define the guardrails and evaluation practices, wire it into CI/CD, and get engineers to trust it and use it.

As a private pilot, I think about reliability differently. Aviation doesn't forgive sloppy design or heroic workarounds — it demands systems that work predictably when people are stressed and time is short. Elegance isn't decoration; it's what makes a system reliable.

Agentic AI

Building with AI before it was a hiring buzzword

Early adopter

Hands-on ahead of the mainstream

Adopted Cursor (2024) and ChatGPT early, finding real production uses and bringing AI-assisted development into everyday engineering practice well before it became an industry category.

Shipped

Proactive failure-triage classifier

Built an AI classifier that categorized failures by type, routed each to the right engineer, and suggested next steps right in the Slack alert to get them moving. Often caught and resolved issues before users were even affected — with a roadmap to auto-generate fixes and update code for new dependency versions.

Cross-team POC

Natural language → SQL

Led a cross-team proof-of-concept using GPT-3.5 to turn plain-English questions into SQL, giving non-engineers self-service access to the data they needed without waiting on engineering.

Production

LLM tooling on Amazon Bedrock

Architected and shipped AI-powered internal developer tools on Amazon Bedrock that automated monitoring, compliance, and engineering workflows — in a regulated, HIPAA-compliant environment.

Projects

Agentic AI, shipped — not just slides

Mealr

A production serverless platform with conversational AI over your own recipes
Solo architect & builder

Upload a scanned recipe PDF and a pipeline extracts it into structured data; then ask plain-English questions ("what can I make tonight?", "add the Tuesday pasta to my list") and a retrieval-grounded assistant answers using only your recipes, with citations. Eight independently deployed services, all infrastructure-as-code.

  • Conversational RAG — Amazon Bedrock Converse + Knowledge Base (Claude Sonnet 4.6), multi-turn chat.
  • Per-user retrieval scoping — answers are provably limited to the caller's own data.
  • Grounded citations — recipe references validated against retrieval; no hallucinated sources.
  • Vector search — Bedrock Knowledge Base on Amazon S3 Vectors, synced via EventBridge.
  • Document-AI pipeline — scanned PDFs → structured JSON via Step Functions + EventBridge.
  • Whole platform — Cognito OIDC/MFA, API gateway, CDK IaC, OpenAPI, pytest, SemVer.
Amazon BedrockS3 VectorsClaude Sonnet 4.6RAGAWS LambdaStep FunctionsEventBridgeDynamoDBCognitoAWS CDKReactTypeScript

A personal product, currently in private beta. Source is private; the case study walks through the architecture and the agentic-AI design.

Experience

18 years building reliable systems

Senior Staff Engineer — AI Platform & Developer Enablement

Mar 2023 – Sep 2025
CareJourney by Arcadia

Drove early agentic AI adoption across engineering, built a CI/CD failure-triage agent and a natural-language-to-SQL POC, and took deployments from monthly to multiple per day — while leading a company-wide shift-left and setting architecture for new applications under HIPAA.

Senior Cloud & DevOps Engineer

Sep 2021 – Feb 2023
CareJourney by Arcadia

Built the CI/CD and infrastructure-as-code foundation (Terraform, CDK, AWS) that later enabled multiple-deploys-per-day, and partnered across engineering and security to make compliance automatic rather than a release-time bottleneck.

Cloud & DevOps Engineering Leader

Jan 2016 – Jun 2021
Datometry

Worked across engineering, product, and customer teams to design cloud engineering efforts across AWS, Azure, and GCP for multiple enterprise SaaS products — architecting infrastructure, driving significant cloud-cost savings, and improving developer productivity through testing and observability.

DevOps Team Lead

Oct 2013 – Sep 2015
Millennial Media

Led the DevOps team for a highly available digital ad service, directing automation tooling for testing and deployment and improving release speed through efficient CI/CD practices.

DevOps Engineer · Library of Congress

2012 – 2013

Earlier: Systems Engineering & Administration roles

2008 – 2012

A deep infrastructure foundation — virtualization, storage, monitoring, and high-availability systems — at Driven, Cooper Notification, and Symplicity.

Toolbox

What I work with

Agentic AI & LLMs

Claude CodeCursorAmazon BedrockLLM integrationEvalsPrompt & context engineering

Platform & Delivery

Developer enablementCI/CDGitHub ActionsCircleCIGitOpsContinuous DeliveryObservability

Cloud & Infrastructure

AWSAzureGCPTerraformAWS CDKKubernetesServerlessDocker

Languages & Practices

PythonNode.jsReactBashSoftware architectureShift-leftHIPAA compliance
Contact

Let's make agentic development real at your company

I'm looking for a Staff/Principal or platform-lead role where I can set architecture for new applications and bring agentic development practices into the org — the kind of change that sticks.