Skip to content

Engineering

Backend, data infrastructure,
and the product problems in between.

I build across product and infrastructure. For seven years, I’ve led platform initiatives and owned critical backend and data systems through architecture, adoption, and scale.

At Udaan, I owned SQL infrastructure across reliability, observability, cost, and developer enablement, helping cut database spend by 75%. I created snorql, Udaan’s first open-source project, and co-founded SIGDB, its organisation-wide database community. As a founding engineer on Percept Insight, I owned the entire query layer and Druid infrastructure, and led performance work as it grew from an internal platform into SaaS.

Earlier, I was a founding engineer at Clarisights, building a high-scale marketing analytics platform.

Experience

Companies I've worked with.

Udaan

Software Engineer

May 2020 — Sep 2024 · Bengaluru

India's largest B2B commerce platform. Owned SQL infrastructure and helped build Percept Insight — an in-house product-analytics platform that grew into a SaaS offering — working across data infrastructure, OLAP systems, and supply-chain tooling.

  • Druid
  • SQL Server
  • PostgreSQL
  • Kafka
  • Python
  • Go
  • Kubernetes
  • OLAP
  • RAG
More about this role Show less
  • → Reduced SQL database costs by 75% over two years — millions in savings — leading SQL infrastructure across reliability, observability, and developer enablement.
  • → Optimised 4+ self-managed Druid clusters: 50% lower query latencies and 30% less resource provisioning via ingestion and read-path improvements.
  • → Founding team member of Percept Insight: built query-constructor-service (central Druid SQL generation for slicing, funnels, cohorts) and percept-interact, an LLM/RAG engine for natural-language dashboards.
  • → Created snorql — a SQL diagnostics & monitoring tool and Udaan's first open-source project; co-founded the #SIGDB database guild.
snorql — blog ↗ snorql — talk ↗

Clarisights

Software Engineer

Jul 2017 — Apr 2020 · Bengaluru

Founding team (<10 people) of a marketing / business-intelligence platform with HTAP workloads for data-driven marketing teams. Full-stack contributor across UI, product, data pipelines, infrastructure, and backend.

  • Ruby on Rails
  • PostgreSQL
  • MongoDB
  • Redis
  • Kafka
  • Kubernetes
More about this role Show less
  • → Implemented tenant isolation across MongoDB clusters and ran a zero-downtime live data migration.
  • → Tuned and managed a master–slave PostgreSQL setup holding 4.5 TB+ of data.
  • → Designed Kafka-based ingestion pipelines (Adjust Analytics) deployed on Kubernetes, and built ETL for ad-channel connectors.
  • → Rebuilt the marketing analytics dashboard (v2) and real-time collaboration via Rails ActionCable.

Celebal

Full-Stack Developer, Intern

Nov 2016 — Feb 2017 · Jaipur

Built a prototype hyperlocal food-delivery application end to end.

  • Go
  • Gorilla/mux
  • CockroachDB
  • Angular

Wheelie Repairs

Co-founder

2016 — 2016 · Jaipur

Early startup building an online platform for doorstep two-wheeler servicing — booking, repairs, and maintenance brought to your door. Co-founded while still in college.

Projects

Independent projects.

A few projects that emerged from curiosity and the desire to create.

MastiDB

An on-disk OLAP database engine built from scratch in Python.

  • Databases
  • OLAP
  • Python
Explore

dhol.dance

A creative tool that makes beat-making intuitive for people who do not think like music producers.

  • Creative tools
  • Music
  • Product design
Explore

Raft Bare Bones

An educational implementation of the Raft consensus algorithm in Go.

  • Distributed systems
  • Go
  • Raft
Explore

Stack

Technologies I work with.

Primary

  • Python
  • Go
  • SQL
  • SQL Server
  • PostgreSQL
  • Druid
  • Kafka
  • Redis
  • Kubernetes
  • Docker
  • OLAP workloads

Worked with

  • Ruby
  • Ruby on Rails
  • Java
  • Kotlin
  • MongoDB
  • GCP
  • Azure
  • Ansible
  • PromQL
  • RAG systems

Talks

Conferences I've spoken at.

Rootconf · Jul 2023

Squeezing the juice out of SQL Server

Udaan's infrastructure cost-optimization journey — focused on SQL Server and CosmosDB, with the tips and learnings behind cutting database costs by 75%.

Slides ↗ Video ↗

InOut 6.0 · Oct 2019

Productionizing Hello World

Taking a simple service all the way to production with Go, Docker, Kubernetes, and an observability stack.

Slides ↗

Technical writing

Notes from the engineering archive.

Older and newer essays on databases, computer science, tools, and the engineering habits behind durable systems.

All writing →
Dancing with heuristic algorithms Engineering · 14 Mar 2023 · 10 min Dancing with heuristic algorithms How to solve NP-hard problems using local search heuristic algorithms — and how CS knowledge can help with real-life… Read essay → SQL Server: Dropping Indexes in production (safely) Engineering · 03 Jun 2021 · 5 min SQL Server: Dropping Indexes in production (safely) Under the hood of dropping an index and the locking mechanism. It also talks about how you can delete indexes safely. Read essay → Unused index analysis in SQL Server with master-slave setup Engineering · 01 Jun 2021 · 4 min Unused index analysis in SQL Server with master-slave setup Identifying Unused indexes using SQL server DMVs involving dm_db_index_usage_stats in a replication setup Read essay →

Begin a conversation

Get in touch.