Skills
Tools I reach for
under pressure.
Split by where it was earned: what has shipped to production at TCS, and what was built self-directed. Every line names the work that evidences it, so the two are not read as one.
Proven in production at work
Shipped and operated at Tata Consultancy Services, Jan 2025 – Present
- PythonFastAPIREST
- 13+ APIs on a production system serving 10,000+ requests/day
- SQLData pipelinesAuthN/AuthZ
- Query and pipeline optimization under production load; enterprise access-control standards
- RAGVector searchEvaluation
- Relevance tuning plus the automated eval harness that measured it
Proven outside work
Designed, built and benchmarked end-to-end in the projects on this site
- RustC++
- Allocation-free feed handler at 1M+ msg/sec, matching engine, FIX 4.4 session layer, pre-trade risk checks
- GogRPCRedisPostgreSQL
- Consistent hashing, Raft leader election, k6 load testing at 45,000 req/sec
- DockerKubernetes (AKS)PrometheusGrafana
- Containerized deploys, metrics, and operator dashboards
- TypeScriptReactNext.js
- This site — designed, built and deployed solo
- ConcurrencySystems
- Lock-free and zero-allocation hot paths, binary wire protocols, distributed consensus
What connects to what.
Laid out by a force simulation run to completion on the server, so it ships as plain SVG and is identical for every visitor — no physics loop on your main thread.
Everything, indexed.
The full inventory, including what is used lightly. Depth is the section above; this is coverage.
24 skills, grouped.
Languages
- C++
- Python
- Go
- Rust
- JavaScript
- TypeScript
Frontend
- React.js
- Next.js
Backend
- FastAPI
- REST APIs
- Microservices
- Distributed Systems
Cloud & DevOps
- Azure Kubernetes Service (AKS)
- Docker
- Kubernetes
- Prometheus
- Grafana
AI/ML & Generative AI
- AI Agents
- Prompt Engineering
- RAG (Retrieval-Augmented Generation)
- Vector Databases
Core CS
- System Programming
- Multithreading
- Concurrency
