Experience

TrueFoundry

• United States
• Software Development
• 1 - 100 Employee

• Co-Founder & CEO

  • Sep 2021 - 2 years 4 months

  San Francisco Bay Area Founded in 2021 & backed by Sequoia, Naval Ravikant, TrueFoundry is a developer platform that enables a company to deploy and monitor machine learning models at the speed of big tech companies (Facebook, Google, Apple) reducing their AI time to value. We aim to unlock business value by enabling teams to run quicker experiments & deploying models without knowledge of complex infra. In addition, we also work closely with CoE teams as their co-pilot around cost reduction & revenue… Show more Founded in 2021 & backed by Sequoia, Naval Ravikant, TrueFoundry is a developer platform that enables a company to deploy and monitor machine learning models at the speed of big tech companies (Facebook, Google, Apple) reducing their AI time to value. We aim to unlock business value by enabling teams to run quicker experiments & deploying models without knowledge of complex infra. In addition, we also work closely with CoE teams as their co-pilot around cost reduction & revenue optimization programs by helping them adopt Data Science & ML. The goal is to enable companies to get better ROI on their ML initiatives. PS - We also allow you to host LLMs on your own cloud/On-prem infra Show less



EntHire

• United States
• IT Services and IT Consulting
• 1 - 100 Employee

• Co-Founder & CEO

  • Jul 2020 - Aug 2021

  United States As engineers, we all know how difficult it is to hire tech talent. So I quit my job and decided to solve this problem with my friends from IIT KGP. Ran EntHire.co for a year before selling it to Info Edge. Read more about the acquisition - https://www.bigshyft.com/blog/bigshyft-acquires-enthire-a-pre-interview-platform With the same founding team, we are now solving a problem I saw very closely while building EntHire.co. Idea is for Data Scientists & MLEs to productionize models without… Show more As engineers, we all know how difficult it is to hire tech talent. So I quit my job and decided to solve this problem with my friends from IIT KGP. Ran EntHire.co for a year before selling it to Info Edge. Read more about the acquisition - https://www.bigshyft.com/blog/bigshyft-acquires-enthire-a-pre-interview-platform With the same founding team, we are now solving a problem I saw very closely while building EntHire.co. Idea is for Data Scientists & MLEs to productionize models without knowledge of complex infra & with the speed of how a large large company like a Facebook or Amazon does it. Ultimately, the goal is to unlock business value. Inspiration: My time working with Facebook where I used the best tools to productize models + how challenging and expensive it is to find good MLOps talent who can build a ML Platform in house :) Show less



Facebook

• Software Development
• 700 & Above Employee

• Machine Learning Tech Lead

  • May 2018 - Sep 2019

  Menlo Park - Working in conversational AI developing Facebook's virtual assistant- Portal. - Worked on launching FB's first on-device model on FB messenger. - Tech Led the development of the Proactive Assistant! - Implemented new infrastructure to setup an event driven pipeline for Facebook Virtual Assistant.



Reflektion

• United States
• Software Development
• 1 - 100 Employee

• Machine Learning Tech Lead

  • Jul 2016 - May 2018

  San Mateo -  Envisioned, architected and jointly implemented AI platform that brings together all machine learning at Reflektion and exposes them to the consumers as smart tools, APIs and microservices. - Orchestrated the launch of recommendation engine on merchandising control center acting as the primary revenue source of the company for 2018. - Implemented a scalable and confi gurable real time prediction layer scaling to a 100x catalog size. - Hired and mentored 4 engineers… Show more -  Envisioned, architected and jointly implemented AI platform that brings together all machine learning at Reflektion and exposes them to the consumers as smart tools, APIs and microservices. - Orchestrated the launch of recommendation engine on merchandising control center acting as the primary revenue source of the company for 2018. - Implemented a scalable and confi gurable real time prediction layer scaling to a 100x catalog size. - Hired and mentored 4 engineers (including cross-geo) to grow the team.

• Machine Learning Software Engineer

  • Feb 2015 - Jul 2016

  San Mateo - Developed the algorithm, architected, implemented and deployed an incrementally updatable product recommendation model resulting in CTR improvement upto 500 %. - Implemented the online evaluation framework for recommendation quality measurements which correlated well with CTR, making development and model tuning cycles much faster. - Responsible for a company wide database migration and adoption of new technologies like Spark. - Refactored clustering pipeline to make it… Show more - Developed the algorithm, architected, implemented and deployed an incrementally updatable product recommendation model resulting in CTR improvement upto 500 %. - Implemented the online evaluation framework for recommendation quality measurements which correlated well with CTR, making development and model tuning cycles much faster. - Responsible for a company wide database migration and adoption of new technologies like Spark. - Refactored clustering pipeline to make it extensible to new models, confi gurable and modular. - Implemented a framework to support client requests for promotional and manual merchandising.



University of California, Berkeley

• United States
• Higher Education
• 700 & Above Employee

• Graduate Student Researcher

  • Aug 2013 - Dec 2014

  I worked with the Design, Modeling and Analysis group at Berkeley, on the development of tools and methodologies for the design of Cyber Physical Systems. My focus was on formal methods and optimization based techniques for the correct-by-construction synthesis of system architectures. Our current efforts include automatic synthesis of reliable architecture of Aircraft Air Management System and Aircraft Electrical Power System.



University of California, Berkeley

• United States
• Higher Education
• 700 & Above Employee

• Visiting Student Researcher

  • May 2012 - Jul 2012

  Developed an algorithm for the automatic generation of optimal single line drawing of Aircraft Electric Power System(EPS). Automated the probability computation of dependent events, solving the reliability computation of DAGs. Framed the architecture synthesis as an optimization problem and developed a tool which is currently being used as the EPS design infrastructure at UC Berkeley.



Technion - Israel Institute of Technology

• Israel
• Higher Education
• 700 & Above Employee

• Visiting Student Researcher

  • May 2011 - Jul 2011

  Implemented an untrained algorithm to detect salient object from any given image. Reduced the runtime of the existing code from 2.75 sec/Image to 1.5 sec/Image using Matlab and OpenCV Library. Validated our results on the MSRA database of 5000 images obtaining results with 94 % accuracy