Engineering Manager, Blob Storage @ Meta
Leading the engineering teams behind Meta's Data Transfer Infrastructure — moving exabytes reliably, one byte at a time.
15+ years · ex-AWS Engineering Leader · org design is just distributed systems with humans
Career
From Yahoo! Bangalore to Amazon to Meta — always deep in distributed systems, always building teams that build infrastructure that cannot fail.
Click any card to read more
Jun 2025 – Present
Meta · Seattle, WA
Leading the team building Meta's exabyte-scale data transfer infrastructure. My org owns the plumbing that moves data in and out of Meta's global systems — ingestion from production services, large-scale exports feeding ML training pipelines, and reliability systems that operate at the edge of what distributed infrastructure can do.
Jul 2022 – Jun 2025
Amazon Neptune · AWS · Seattle, WA
Led four of Neptune's seven teams through the database's highest-profile era. My org shipped Neptune Analytics at re:Invent 2023 and co-delivered GraphRAG with Amazon Bedrock at re:Invent 2024. Rebuilt the on-call model from the ground up — tiered rotation, follow-the-sun coverage — and drove engineer satisfaction from 50% to over 90%.
Jan 2020 – Jul 2022
Amazon Neptune · AWS · Vancouver, BC
Built the Neptune Storage team from one engineer to fifteen — across Seattle, Vancouver, and the Bay Area — while simultaneously scaling Neptune from 6 to 22 AWS regions. My team delivered the NVMe cache layer that drove 30% query speedup and ~$4M in incremental ARR, and cut full-graph S3 exports from days to under three hours.
Oct 2011 – Dec 2019
Amazon · AWS · Seattle, WA
Eight years as IC and tech lead — the foundation for everything that came after. As tech lead, I owned Neptune's Control Plane architecture, drove a TLS overhaul adopted by four other AWS services, and built the core enterprise feature set from scratch. Those systems have been in production across every Neptune version since.
Summer 2011 · 3 months
Yahoo! Bangalore · Bangalore, India
Three months at Yahoo!'s Bangalore engineering center, working on ad quality pipelines — malware detection signals to catch bad ads before they served. First real exposure to production systems at scale. Graduated from NIT Calicut a few weeks later and joined Amazon.
Jun 2007 – May 2011
National Institute of Technology, Calicut · India
What I bring
I've scaled engineering orgs from 1 to 30+ across multiple geographies. Not by adding headcount — by structuring work so each engineer multiplies the team's output. The metric I optimize for is org velocity, not individual throughput.
I push teams toward systems that become platforms. Neptune's Control Plane became the shared foundation for DocumentDB, Timestream, and MemoryDB. Good infrastructure has a multiplier effect — I build the infrastructure that lets other teams build faster.
Every major call has a model behind it. On-call redesign moved satisfaction from 50% to 90% — I measured it first. Region expansion became metadata-driven, zero code changes per region — I designed for that scale before we needed it.
I give each engineer their own problem space — the customer context, the success metric, and the keys. In the agent era that means enabling every person to run a squad of AI agents that multiplies their output. Mini-startups inside a large org ship things large orgs can't.
Ship a working v1 in days, not quarters. The prototype teaches you more than the spec. I build teams comfortable with rough edges — because when agent-assisted development makes prototypes nearly free, the winning move is to put something real in front of users first.
Shipped
Fourteen years shipping at Amazon and Meta. These are the ones that moved the needle.
Click any card to read more
Founding SDE / Tech Lead
AWS's first fully managed graph database — built the Control Plane from scratch.
GA at re:Invent 2017 · 30K+ servers globally
SDM, Storage & Control Plane
New graph analytics engine — built four teams' worth of it, shipped at re:Invent 2023.
Launched re:Invent 2023 · GA 2024
SDM, Neptune + Bedrock
Graph-powered retrieval-augmented generation — shipped live at re:Invent 2024.
GA · Demoed live at re:Invent 2024
Tech Lead, Neptune
Zero-trust authentication and hardened TLS for Neptune — adopted across AWS.
Architecture adopted by 4+ AWS services
SDM, Neptune Analytics Storage
New in-memory ETL model for Neptune Analytics — 10M+ records/sec.
80x faster than Neptune DB · 10M+ records/sec
SDM, Neptune Storage
NVMe-backed buffer pool for hot graph data — 30% query speedup.
30% query speedup · ~$4M incremental ARR
Blog
Side Projects
Vibe coded on free time — mostly to make my own life easier. Claude at the core, shipped to prod because why not.
AI-powered options trading copilot. Live market data, PMCC scanner, real signals — built with Claude + Yahoo Finance. Not financial advice, definitely financial vibes.
Real-time sports club scheduling and court management. AI-assisted rotation and matchmaking — an engineering marvel the club didn't know it needed.
AI-powered tax research assistant. Ask complex questions, get reasoning-backed answers. Built with Claude for workflows where accuracy actually matters.
108 original superhero stories across 20 heroes — AI-generated, fully static. Marvel, DC, and Avengers universes. Built for my kids.
GitHub org for weekend vibe-coding sessions. Real tools, real data, real AI — zero stakeholders, zero meetings. This is what happens when an infrastructure engineer builds product.
View on GitHub