About

Mon, 01 Jan 0001 00:00:00 +0000

👋🏻 Hi, I’m Omkar.

I’m a software engineer based in Pune, India. I like building distributed systems, data pipelines, and AI agents that work in the real world — not just in demos.

I graduate from Pune Institute of Computer Technology in June 2026 with a B.E. in Electronics & Telecommunication.

What I think about #

A lot of my work sits at the boundary between metadata, data infrastructure, and durable systems — connectors, lineage extraction, event-driven pipelines, and the kind of glue that has to keep working when things go wrong. Some themes I keep coming back to:

Projects

Mon, 01 Jan 0001 00:00:00 +0000

A selection of things I’ve built.

AI #

sentinel-agent — AI-powered GitHub PR-review agent. Temporal for durable orchestration, LangGraph for multi-step reasoning, Neo4j knowledge graph (CALLS / IMPORTS / CONTAINS), Tree-sitter AST extraction, validator stage to ground every finding. p90 latency under 5 min, 95% review-publish success. Python 2026
echo.ai — Real-time audio bridge between developers and AI coding agents. Narrates what the agent is doing, alerts when it’s blocked, accepts voice responses — so you never have to return to the screen. Python 2026

Work Experience

Mon, 01 Jan 0001 00:00:00 +0000

Atlan — Software Engineer I #

Jan 2025 — Present · Apps Development · Pune, India (Remote, SF HQ)

Owner of metadata pipelines across Snowflake, Databricks, BigQuery, dbt, Kafka, and Matillion on Atlan’s connector platform. Subject Matter Expert for the Tags pipeline. Stack: Python, Java, Go, Kubernetes, Temporal, Argo, Kafka, Neo4j.

Highlights

Snowflake Listings & Shares — Engineered ingestion end-to-end (typedefs, ETL, lineage) powering Atlan Data Products. Demoed at Snowflake Summit as a flagship capability.
Snowflake Secure Agent — Architected Crawler/Miner to extract metadata inside the customer’s Kubernetes cluster without sharing credentials. Unblocked enterprise deals at JPMC, Morgan Stanley, and P&G.
Databricks Cross-Workspace Lineage — Owned end-to-end across multi-workspace edge cases (masked entities, private catalogs, system-table joins). Positioned Atlan as the differentiated source of truth over native Databricks UC (which renders cross-workspace assets as masked objects).
Databricks OAuth/SSO + Cross-Workspace OAuth — Engineered for Insights and Sample Data preview. Unblocked P&G and Kaizen Gaming.
Databricks lineage generator — Engineered graph-based generator for PATH-type and temp-table lineage; collapsed multi-hop temp-table chains into clean upstream-downstream edges. Reduced major tech debt and improved lineage observability for every Databricks customer.
Unified Tag Contract — Owned the Tags Sync Pipeline across Snowflake, BigQuery, Databricks, and dbt as SME. Architected a unified contract consolidating three divergent connector flows into one publish-app implementation. Architected tag drift observability and reconciliation via the Metadata Lakehouse.
Kafka V3 + Temporal Native migration — Owned migration of all three flavors (Apache Kafka, Confluent, Amazon MSK) to App SDK V3 + Temporal Native Orchestrator, retiring Argo workflows for the connector.
Customer escalations — Drove on-call resolution for easyJet, GM, Loopback, AIMCO, ASOS, Virgin Media O2, FIFA, Dropbox, NYL, Sargento, Tide. Shipped permanent fixes with sub-24hr turnaround on RCAs.
Performance — Optimized SQL queries and I/O to handle 62M+ records; reduced Kubernetes pod memory footprint by ~10GB (Parquet + DuckDB processing), eliminating recurring OOM failures.
Matillion Custom CA Cert support — Engineered SSL custom-CA import flow; saved Sargento’s PrivateLink integration when the customer was about to abandon.

Shelf (YC S22) — Backend Developer Intern #

Aug 2024 — Oct 2024 · Bengaluru, India (Remote)

Home on Omkar Gade

About

What I think about #

Projects

AI #

Work Experience

Atlan — Software Engineer I #

Shelf (YC S22) — Backend Developer Intern #