Vasu Lakhani

About

Backend/AI workflow engineer with experience at AirKitchenz, Uber, and Vivma Software, building production systems on AWS (Lambda, DynamoDB, Step Functions). Has a track record of major performance wins (DynamoDB latency 2s to <150ms; Postgres query 2s to ~180ms) and shipping LLM-powered onboarding and ticket-routing workflows with strong guardrails (schema validation, confidence thresholds, human-in-the-loop escalation).

Experience

Education

Key Strengths

• Diagnosed and eliminated DynamoDB scan-driven latency via schema redesign (composite keys, GSI) and Redis caching (2s to <150ms)
• Built and shipped LLM-assisted onboarding with validation/enum checks and confidence-based human review (cut onboarding time ~40%)
• Designed reliable multi-step agent/workflow orchestration with AWS Step Functions, tool selection, retries with exponential backoff, and SQS-based escalation
• PostgreSQL performance tuning using EXPLAIN ANALYZE, composite indexes, query rewrite, and Redis caching (2s to ~180ms)
• Production Python engineering practices: layered architecture, typing (mypy), data validation (pydantic), dependency management (poetry), and testing (pytest)
• Built real-time, decision-driven backend services processing continuous event streams
• Strong focus on correctness, low latency, and fault handling in production systems
• Debugged latency/state issues using logs, metrics, and event-flow replay
• Reduced downstream errors by 30%+ by adding stricter validation and fallback handling
• Designed reliable inter-service communication with validation, retries, and monitoring
• Production deployment experience with Docker on AWS ECS and CI/CD via GitHub Actions