Himanshu Baliyan
Software Engineer — Backend & AI Systems
himanshubaliyan4000@gmail.com+91 70068 02968linkedin.com/in/himanshu-baliyanGurugram, India
Summary
I build production backend services in Java and Spring Boot, and ship data and LLM systems end to end: forecasting pipelines, RAG, agentic workflows and the plumbing underneath them.
I care about the unglamorous parts — p95 latency, pipeline uptime, tracing coverage, deploy time — because those are what decide whether a feature survives contact with real traffic.
Experience
Software Engineering Intern — Airtel Digital
- 6 services / 25+ endpoints. Built and maintained backend microservices and REST endpoints in Java, Spring Boot and PostgreSQL, supporting enterprise analytics workflows used by 4 internal teams.
- 2 hours → under 5 minutes. Developed AI-powered backend workflows with LLM APIs, LangChain and a RAG retrieval layer, cutting enterprise document lookup per request.
- 12M events / day at 99.5% uptime. Designed distributed Apache Spark and Kafka pipelines ingesting telemetry, reducing batch processing time by 35%.
- 90% monitoring coverage, −40% MTTD. Instrumented 8 services with Prometheus, Grafana, Loki, Tempo and OTLP tracing, cutting mean time to diagnose incidents.
- 45 min → under 10 min releases. Automated containerised deployments with Docker, Kubernetes and CI/CD, cutting failed deployments by 60%.
- 850 ms → 300 ms p95. Optimised SQL queries and service APIs under production load.
Selected projects
Air Clear — Air-Quality Outlook for Schools
Hourly ingestion from three sources into TimescaleDB, orchestrated by nine Airflow DAGs. Each of the next five days is graded on the official CPCB scale and served by a FastAPI API to a React dashboard, with a watchdog, tested backups and a nightly evaluation behind it.
Result: Live for about 80 stations. In backtests the grade is exactly right on about 6 in 10 days for tomorrow, and a missing forecast is never shown as "go".
Agentic Job Automation Platform
Graph-based multi-agent orchestration with tool calling and structured memory, persisting state between agent steps so a run can be resumed and inspected.
Result: End-to-end task execution instead of one-shot suggestions.
Skills
- Languages
- Java · Python · SQL · TypeScript · JavaScript · Bash
- Backend & Data
- Spring Boot · FastAPI · REST APIs · Microservices · Apache Airflow · Apache Spark · Apache Kafka
- Forecasting & ML
- pandas · scikit-learn · LightGBM · Time-series backtesting · Model evaluation
- AI & LLMs
- RAG pipelines · LangChain · LangGraph · Agentic workflows · Tool calling · MCP · Embeddings & vector search · Prompt engineering
- Databases
- PostgreSQL · TimescaleDB · MongoDB · OracleDB
- Cloud & Ops
- AWS · Docker · Kubernetes · Linux · CI/CD · GitHub Actions · Cloudflare Workers · Prometheus · Grafana · Loki · Tempo · OpenTelemetry
Education
Master of Computer Applications