• Currently: Founder @ Autonmis AI-native governed data platform for Ops & Business.
• Previously: AVP Data & ML @ Paytm · Head of Data @ Junglee Games, Lido, Simplilearn
• Built multi petabyte scale data processing platforms at Paytm & Junglee Games
• Focus: Data platforms, Agents, distributed compute systems, governance, AI execution layers

• System design: https://www.linkedin.com/in/abhranshu-bagchi/recent-activity/articles/
• Case studies: https://autonmis.com/blogs
• Articles: https://medium.com/@abhranshubagchi

Designing AI-native governed execution systems on enterprise data

@Autonmis Full-stack AI-native data platform built largely solo. Microservice led multi-tenant org architecture with RLS and 7-year audit logging. Governed access & metric lifecycle (defined → draft → published) enforced across dashboards, alerts, and executive briefs. LangGraph session graphs + BullMQ worker pools + OpenRouter model routing with fallback chain. On-premises EKS/EC2 deployment with VPC isolation for regulated customers. Early deployment on customer data in diverse verticals.

@Paytm — Data & ML Platform: Ledger reconciliation and financial reporting across ₹250Cr+/month in flows. MLOps for AML in Payments. Payout engine for 100K field agents — rules engine, scoring system, device servicing workflows. Governance and observability layer across 1,000+ production pipelines.

@Junglee Games — Data Eng & ML Systems: Real-time gameplay ecosystem at 1B+ daily events. Sub-second leaderboard updates for 10M concurrent players. Financial-grade transaction processing for 50M+ monthly transactions. Data org built 0 → 22 engineers.

Most of my work is in private repositories.

This repository captures:

• system design decisions
• architecture patterns
• lessons from building production data systems

I take on a small number of engagements where the problem is hard and the context fits.

Architecture sprint — 2–4+ week embedded engagement. Current-state audit, architecture decision document, implementation roadmap. For teams at the inflection point between scrappy and engineered.

Platform deployment — Full Autonmis deployment on your infrastructure, configured for your data and governance requirements. For lean teams that need the system working, not just designed.

Fractional data leadership — Ongoing technical anchor for data decisions: architecture reviews, hiring, roadmap. For companies between "no data leader" and "ready to hire full-time."

Book a call: https://cal.com/abhranshu-b/exploratory