Flagship engineering
Systems I've built and shipped
Selected engineering 9 systems
01 Data connector IP, now in AWS Athena Designed federation connectors for Snowflake, BigQuery, IBM DB2, and Oracle that AWS acquired and integrated into Athena - and worked directly with AWS engineering on the transition. Acquired by AWS 02 CMigrate, a Rust block-level replication engine A Rust agent that live-migrates on-prem servers to the cloud by streaming disk blocks with checksum-based incremental scans - coordinated multi-disk cutover, durable resume, near-zero downtime. Live servers, near-zero downtime 03 Concierto Scout, an IT discovery appliance An agentless appliance that ships as a single VM image and maps a customer's estate - servers, databases, performance, and dependencies - over SSH, WinRM, and TDS, with nothing installed on targets. Agentless discovery 04 Maximize, an AI FinOps agent for AWS & Azure A multi-tenant FinOps platform that turns daily billing feeds into resource-level anomaly signals and concrete optimization actions - rightsizing, cleanup, Graviton, commitments - with an executive dashboard and chat. Cost anomalies → actions 05 A multi-agent bug-fix orchestrator A pipeline of analyser, fixer, validator, and git agents - with human approval gates and workspace isolation - that takes engineers from bug report to merged PR in under 20 minutes. Bug → merged PR in <20 min 06 An autonomous PR review agent Reviews every pull request end-to-end - code quality, architecture, security, test coverage - with inline comments and verdicts at zero human kickoff. Manual review cycles down 60%+. Hours → minutes 07 Enterprise RAG at production scale Hybrid dense + sparse retrieval with cross-encoder reranking across 50K+ enterprise documents. Top-5 retrieval precision lifted from 72% to 94%; latency cut from 4.2s to sub-800ms. 96% accuracy · <800ms 08 Private AI coding CLI on self-hosted Gemma 3 A terminal-native developer assistant on in-house GPUs - codebase-aware retrieval, tool use, multi-step execution - keeping AI-assisted development entirely inside the enterprise security perimeter. Zero external calls 09 AI-native SDLC across the organization AI agents integrated across requirements, architecture, coding, testing, and deployment, in use by all five product clusters - plus 6+ production business agents spanning FinOps, claims, knowledge search, and migration. Adopted by all five clusters