Senior Software Engineer | Python | Golang | Java | Agentic AI | Microservices | Distributed Systems | Multithreading & Concurrency | Django | FastApi | Gin | SpringBoot | AWS | GCP | PostgreSQL | MongoDB | Milvus

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Noida, Uttar Pradesh, India

Noida, Uttar Pradesh, India







Bachelor of Technology, Information Technology
Institute of Engineering and Technology
Join Tarun and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossSenior Software Engineer | Python | Golang | Java | Agentic AI | Microservices | Distributed Systems | Multithreading & Concurrency | Django | FastApi | Gin | SpringBoot | AWS | GCP | PostgreSQL | MongoDB | Milvus
I build backend systems and AI agents that ship to production. Right now, I'm a Senior Software Engineer at Impact Analytics, building the agentic rebuild of AssortSmart, a merchandise planning SaaS where enterprise retailers decide what to buy, how much, and for which stores. AI agents draft store clusters and plans, and planners review, override, and approve. My work spans the Python agentic microservice (FastAPI, LangGraph, MCP) that runs all agentic workflows, the Go (Gin) core backend, and the data layer: ClickHouse with an append-only versioned write model where every planner edit is an insert, giving us version diff, undo/redo, and an immutable audit trail for free. Before this I spent two years at Uber (via EPAM Systems) on two very different problems: a financial risk-scoping platform for Uber Finance that cut a quarterly manual cycle by 70% and feeds PwC audit work papers, and the Uber Eats menu ingestion platform, an event-driven pipeline (Kafka, Flink, Spark) processing 30,000+ vendor menus a month with an AI extraction layer (RAG + Gemini) at 98% fidelity, saving $600K+ annually. Earlier, at Masters India, I led the migration of a GST e-invoicing platform from a PHP monolith to async Python microservices, cutting p95 latency from 1.2s to 300ms while scaling to 1M+ transactions a day. I started my career at GeeksforGeeks, rebuilding backend systems used by 100,000+ daily users. What I care about: systems that stay correct under scale, AI with human approval gates instead of blind autonomy, and boring, readable code that anyone on the team can own. Tech I work with daily: Python, Go, FastAPI, Gin, LangGraph, LangChain, MCP, RAG, ClickHouse, PostgreSQL, Kafka, Flink, Spark, Redis, Docker, Kubernetes, GCP, AWS. Open to senior backend / AI platform roles. Reach me at tm.eng2021@gmail.com.
Bengaluru
claude code
cursor
langchain
langgraph
langsmith
Milvus
Senior Software Engineer at Impact Analytics, May 2026 - Present · 4 mos, Bengaluru, Karnataka, India, Building the agentic rebuild of AssortSmart, a merchandise-planning SaaS where AI agents draft store clusters and assortment plans, and planners review, override, and approve them. • Developing the Python agentic microservice (FastAPI, LangGraph, MCP) that runs all agentic workflows. Its store clustering agent selects features based on statistical significance, optimizes the number of clusters via silhouette scoring, and presents 3 to 5 ranked scenarios for planner approval, cutting turnaround time from days to under 1 hour. • Building the core backend as Go (Gin) microservices: REST APIs for plan lifecycle, tenant configuration, and bulk saves, with JWT auth middleware, goroutine worker pools, and context-based timeouts. • Led adoption of ClickHouse as the planning store with an append-only versioned write model where every planner override is an insert, powering version diff, undo/redo, and an immutable audit trail. • Designing the backend for the planning grid, an Excel-like pivot surface serving rollups and drill-downs from pre-aggregated ClickHouse data at p95 under 500ms, with cell edits under 80ms via async write-back. Skills to attach: Go · LangGraph · ClickHouse · FastAPI · Microservices
Software Development Engineer 2 (External) at Uber, Jun 2024 - May 2026 · 2 yrs, Bengaluru, Karnataka, India, Worked on Uber Finance and Uber Eats platform teams. Financial Risk Management (FRM) Scoping Platform — Uber Finance • Cut Uber's quarterly financial risk-scoping cycle time by 70% by replacing a manual Google Sheets workflow with a FastAPI + MySQL platform serving 36 REST endpoints across 8 screens, with output feeding PwC audit work papers. • Designed an 8-table normalized MySQL schema whose quarterly ETL aggregates 19M+ general-ledger rows into 300K rows, reconciled against public 10-Q filings. • Migrated raw SQL to SQLAlchemy 2.0 ORM, fixing a latent column-aliasing bug and raising changed-module test coverage to 100%. Menu Ingestion & Automation Platform — Uber Eats • Built an event-driven ingestion pipeline (Kafka, Apache Flink, Apache Spark) processing 30,000+ vendor menus per month, cutting onboarding from 24 hours to 2 hours and saving $600K+ annually. • Built an AI extraction pipeline with RAG and Gemini 2.5 Pro converting PDF and image menus into schema-validated catalog data with 98% fidelity. • Increased successful ingestions by 95% via IP rotation and dynamic proxy pools, and built ingestion health dashboards on Apache Pinot. Skills to attach: Python · Apache Kafka · Apache Flink · MySQL · RAG
SDE 2 at Masters India, Dec 2022 - Jun 2024 · 1 yr 7 mos, Noida, Uttar Pradesh, India, GST compliance and e-invoicing SaaS used by 2,500+ enterprise clients. • Led the migration from a PHP (Laravel) monolith to asynchronous Python (FastAPI) microservices, cutting p95 API latency from 1.2s to 300ms. • Scaled bulk e-invoice processing to 100,000+ transactions per import and 1M+ daily transactions with chunked async workers and idempotent retries. • Set up centralized ELK logging and New Relic monitoring with alerting, cutting incident triage time by 70%, and raised test coverage from 35% to 82% with pytest automated tests gating CI/CD releases. • Implemented encryption, RBAC, JWT auth, and audit logging for financial compliance data. Skills to attach: FastAPI · PostgreSQL · Redis · ELK Stack · Celery
Software Engineer at GeeksforGeeks, Aug 2021 - Nov 2022 · 1 yr 4 mos, Noida, Uttar Pradesh, India, • Redesigned and migrated the backend from PHP to Django (Python), improving reliability for 100,000+ daily queries. • Designed data models and REST APIs for voting and pinning features, increasing premium subscription sales by 15–20%. • Developed an influencer dashboard with real-time analytics and async cron pipelines for video processing, boosting course sales by 30%. Skills to attach: Django · Python · MySQL · REST APIs · AWS
Bachelor of Technology, Information Technology, Institute of Engineering and Technology
github: https://github.com/tm-dev-pro