Lead Architect & AI Engineer — AI4Tech @ J.P. Morgan Chase

I build production AI systems at enterprise scale. Currently leading DevGPT — J.P. Morgan AWM's AI coding agent platform, now serving 6,500+ engineers, product managers, and senior leaders, and powering Technology Lifecycle Management (TLM): agentic, automated software maintenance across JPMC's full technology estate.

• DevGPT / AI4Tech — Anthropic + OpenAI agentic coding platform on AWS Bedrock and Azure OpenAI. ~90% prompt cache hit rate. 20M peak cache reads/min. Sub-2s sandbox dispatch via Anthropic SRT on EKS.
• Technology Lifecycle Management (TLM) — Firm-wide agentic software maintenance. Agents patch Terraform, Python, Java, and Golang codebases on maintenance branches. Supervisor agent routes changes to owning teams for human-approved merge.
• AWS re:Invent 2026 — Prospectus submitted: DevGPT Coding Agents and its Effect on the SDLC

Languages      Golang · TypeScript · Python · JavaScript · SQL · Rust*
AI / LLM       Anthropic Claude · OpenAI · AWS Bedrock · Langchain · LangSmith · TruLens · W&B
Infra          AWS (EKS · EFS · Bedrock · Karpenter · Lambda · DynamoDB) · Azure · Kubernetes · Docker · Kaniko · Terraform
Networking     Cilium · Istio Ambient · Envoy · CNI · Bubblewrap
Orchestration  Temporal · Kafka · NATS · KEDA · Airflow
Observability  Datadog · OpenTelemetry · Prometheus · Grafana · Phoenix (Arize)
Frontend       Svelte · React · Next.js · Gin · gRPC · GraphQL

• 📣 FINOS 2025 — Transforming Developer Productivity: Rethinking the SDLC with AI Coding Agents
• ✍️ Towards AI — Beyond the Session: Memory Engineering for Agent Teams
• ✍️ AI Advances — Cloud Ant Colonies · Convergent Engineering: How Everyone Built the Same Thing
• 🌐 jordan-carson.github.io — DevGPT architecture, agentic systems, prompt engineering → Full journey