Sep 2021 – Aug 2022 · 11 mos · Bengaluru, Karnataka, India

Jan 2021 – Present · 5 yrs 3 mos

Jan 2016 – Present · 10 yrs 3 mos · Bengaluru, Karnataka, India

• Over the years I have advised many young companies get to their Series-A/B rounds and occasionally invested in them. These include companies like Fyle, Hasura, Primenumbers.in, NextBillion, RudderStack, Bintix and a few others. Open to DMs from startups needing help on product/strategy/financing/engineering/scaling/hiring.

Aug 2011 – Oct 2020 · 9 yrs 2 mos · Bengaluru Area, India

• Qubole built one of the earliest Big-Data as a Service offerings in the public Cloud and was the first vendor to introduce Auto-Scaling Hadoop/Spark/Presto clusters, seamless integration of Spot instances, Columnar caches and an end to end Data Engineering, Data Analysis, and Machine Learning Platform on all major public clouds (AWS/GCE/Azure/OCI). We raised ~$100M, scaled revenues to $50+M ARR with many Fortune-500 customers before exiting to PE. I ran Engineering and Operations throughout - and later also Product. Also frequently donned architect hat and helped channelize advanced technology efforts. Built, scaled and ran the India office from 0 to ~200 people and worldwide Engineering to ~150. Hired and mentored many rock star engineers and built a technology team known for its culture and cutting-edge work.

Jul 2007 – Jul 2011 · 4 yrs

• I worked on different things at FB, in rough chronological order:
• started Hadoop at FB - installing, writing first set of jobs, populating first set of datsets from Scribe and Mysql for consumption by other engineers.
• conceptualized and built internal prototype for Hive and then co-authored open-source Apache Hive as part of Data Infrastructure team.
• managed Data Infrastructure team (about 10-12 people) for about a year. Collectively - the team shipped open-source Hive in this timeframe as well as an internal workflow engine (Databee), new site instrumentation (Nectar), browser based query authoring tool for Hive (HiPal) and FairScheduler for Hadoop.
• contributor of architecture team for FB Messages - prototyped back-end on Cassandra and HBase and made HBase happen at Facebook.
• contributor to FB Credits back-end - helped build credit tracking system, wrote the reporting back-end for FB credits, set up data warehouse for all FB credits data and worked with Finance and IT to put new reporting tier in production for delivering payouts to developers.
• in my last year - worked on Hive and Hadoop:
• automatic local mode in Hive
• shared cache for Hadoop job jars
• helped port and productionize new versions of speculation and preemption protocols in Hadoop Fair-Scheduler, dramatically increased concurrency and cpu efficiency in Hadoop Job-Tracker/Fair-Scheduler
• designed and co-authored new FB-internal version of scheduler (Corona) that is replacing original Hadoop scheduler.
• aside from direct contributions - helped mentor and guide a lot of people on the team and helped set direction for technology initiatives in the team as a whole.

Sep 2005 – Jun 2007 · 1 yr 9 mos

• Founding Engineer for Yahoo's in-house Recommendation Platform. Took from concept to architecture and design and deployment at Yahoo Travel, Shopping, TV and Real-Estate. Currrently working on prototyping new generation of campaign planning tools for Yahoo's Graphical Ads Business.

May 2002 – Oct 2002 · 5 mos

• 6 months of hell and heaven at a seed stage startup. High Point: Writing a new distributed file system in Linux kernel in 3 months. Low Point: Seeing it all wither away.

May 1999 – Sep 2005 · 6 yrs 4 mos

• Designed and developed many parts of Netapp's high-availability and disaster recovery solutions. Initiated and took from concept to reality ground-breaking ideas on file system recovery and scalability.

Feb 1997 – May 1998 · 1 yr 3 mos