Ankit Panicker

I build production-grade software platforms from initial architecture to live operations, taking end-to-end ownership of architecture, implementation, testing, infrastructure, deployment, observability, and production support. My work sits at the intersection of real-time AI infrastructure and reliability engineering. I build systems that must survive real-world conditions—live phone calls that can't drop mid-sentence, hospital workflows that continue during network failures, and media pipelines where one failing component doesn't bring down the rest. I design for the failure path first and the happy path second. Selected production results: • 15,000+ calls dialed, including 6,800+ interactions handled within a 5-hour window by an autonomous hospital voice platform in production. • Multi-tenant AI SaaS platform with tenant-aware isolation serving production healthcare customers. • Reliability validated through load testing (17,856 requests, ~99 req/s, p95 ≈ 572 ms) and 303/308 automated tests passing. • Delivered 16 documented production-hardening initiatives focused on resilience and operational reliability. Engineering principles: • End-to-end ownership—from React frontend to Kubernetes deployment. • Reliability by design using circuit breakers, idempotency, retries, dead-letter queues, graceful shutdown, WAL replay, adaptive backpressure, and structured observability. • Production-first engineering for systems where downtime directly impacts customers. I use AI extensively to accelerate implementation and automate repetitive engineering work, allowing me to focus on architecture, reliability, and product decisions. My strongest areas are Python/FastAPI, Redis, distributed queue/worker systems, and reliability engineering. I also work with TypeScript/React, PostgreSQL, Firestore, Docker, CI/CD, Kubernetes (AKS), and Terraform. In AI, I focus on integrating STT, LLMs, TTS, and RAG into reliable production platforms rather than training models. I thrive in early-stage and high-growth companies where one engineer is expected to deliver architecture, implementation, deployment, and operational ownership. Live products • AIM Studio SaaS: https://aimstudio.co.in/app • AIM System HMS: https://aimsystem.in/hms • GitHub: https://github.com/mrankitpanicker