May 2025 – Oct 2025 · 5 mos · Mountain View, California, United States · On-site

• Trust & Safety @ YouTube
• Designed a new ML-based model to detect and intervene account takeovers across YouTube
• Architected and led the development of this hijack detection system; boosted recall by over 5x (previous 15% to new 78%) at >98% precision

Jan 2023 – Apr 2025 · 2 yrs 3 mos · Mountain View, California, United States · On-site

• Trust & Safety @ YouTube.
• Received T (Transformational Impact) rating for year 2024 — awarded to 3–4% of Googlers.
• Led a team of 5+ SWE, collaborated across 10+ xfn teams to build a self-service tool that notifies and guides hijacked creators through identifying and reverting unauthorized activities across their channel
• Implemented the cleanup execution using a highly scalable, graph-based asynchronous execution framework, enabling complex, non-linear, and reliable recovery workflows
• Drove entire product lifecycle, from ambiguous requirements to a robust solution that now helps 2k users/month
• Received Engineering Excellence Award for the technical design of this product
• Received YouTube Product Excellence Award (awarded to one project across YT every 6 months)
• Received 3 Engineering Excellence Awards, and 6 spot bonuses

Jun 2022 – Sep 2022 · 3 mos · Mountain View, California, United States

• Worked as a part of the Procella (a db query engine) Batch team in the YouTube Data Warehouse
• Implemented data persistent layer to make batch query executions resilient to server crashes and network disruptions
• Built a prototype for batch server and integrated it in Procella architecture to execute batch queries parallely that reads tera bytes of analytics data

Nov 2019 – Aug 2021 · 1 yr 9 mos · Bangalore

• Part of the company’s 125x growth in valuation from $30M to $3.75B
• Architected and developed end-to-end distributed Limit and TWAP Order Management and Execution System
• Spearheaded development of streaming-based low-latency trading application
• Designed and trained a deep neural network to model cryptocurrency price movement bounds after 5 seconds over 9 exchanges on 1 billion data points; built distributed data pipelines to process and label training data
• Developed a Go-based connection pooling and proxy server; integrated 6 liquidity providers using WebSocket
• Devised predictive RFQ (Request for Quote) wait model using real-time probability density of exchanges’ latency

Jul 2017 – Oct 2019 · 2 yrs 3 mos · Mumbai, Maharashtra, India

• Developed equation-oriented scripting module to define cross relations, publish, and price FX structured products comprising of options, derivatives, forward, swaps, and fixed-interest products
• Built end-to-end Maker-Checker framework for CRUD authorizations on DAOs across the entire database; live in 3 international banks
• Developed real-time event-driven PnL module for foreign exchange traders featured with scenario analysis and tenor bucketing
• Created RFQ (Request for a quote) and Trade booking web services using WSDL for integration with IDFC First Bank's treasury system
• Built a cloud-based machine learning app to automate examination of trade finance documents

Mar 2017 – May 2017 · 2 mos

• Instructed a batch of 50 first-year students on 'Linear algebra' course
• Course content:
• Vector space and subspaces, System of linear equations, Guass elimination, Abstract vector spaces, Linear transformations, Gram-Schmidt process, Orthonormal basaes, Eigenvalues and eigenvectors, Special matrices, Algebraic and geometric multiplicity, Spectral theorem

Jan 2017 – Feb 2017 · 1 mo

• Worked as Secondary Teacher for course of "Ordinary Differential Equations" for 50 first-year students.
• Course content:
• Homogeneous and Non-homogeneous Ordinary Differential Equation, Higher order equations, Cauchy-Euler equation, Wronskian, Laplace transform, Step functions and shifting theorems, The convolution theorem

Sep 2015 – Dec 2016 · 1 yr 3 mos

• Worked as a Teaching Assistant for
• CS101 (Introduction to Computer programming) (twice)
• CS213.1x (Foundation of Data structures)
• CS213.2x (Implementation of Data structures)
• CS213.3x (Algorithms)
• Ideated and created randomized quiz questions and auto-graded programming assignments
• Administered the discussion forum on edX and IITBombayX platform
• 40,000+ students' enrollment all across the nation

May 2016 – Jul 2016 · 2 mos · Ann arbor, USA

• Worked on improving wax deposition model predictions through the formulation of a state-of-the-art mathematical model. Programmed MATLAB-based system for comprehensive data analysis of experiments and validation of the model; reduced 40% working hours
• Developed a learning module for 300+ students on the interaction of cobra-venom and antivenom with receptor sites. Modeled and simulated interaction of venom and antivenom with receptor sites as a catalytic reaction. Demonstrated pharmacokinetics of cobra bite using chemical engineering principles through a 15-minute video.