Jan 2024 – Nov 2025 · 1 yr 10 mos · United Arab Emirates

• (Prev Omniverse)
• 1) Research on survival analysis and censored data for Bid price prediction in RTB Protocol. Improved winning rate-2.2x . Serving @ 100k RPS
• 2)Ad moderation and abuse detection through CLIP to map campaigns(videos, text and images) to an embedding space for in images, text content and short-video ads.
• 3) Stateful RAG agent workflow using GPT4(LLM), CLIP(Query Embedder), ControlNet(Generation model) for campaign artifacts generation based on user preferences & past context supporting multi-modal inputs (Images + text).

Apr 2023 – Nov 2024 · 1 yr 7 mos · Delhi, India · Remote

Nov 2020 – Jan 2024 · 3 yrs 2 mos · London, England Metropolitan Area

• 1) Deployed Real time fraud detection for card payments. Tech stack - kubernetes, fastAPI, Couchbase, apache beam. Reduced False positive rate by 5x.
• 2) Deployed inhouse computer vision to enable auto bill-payments, expense management, split bill @ 1sec latency (99th percentile) ~ generating 6 million GBP ARR
• 3) Replaced traditional text detection+recognition models with state of the art vision transformer models( SWIN+ BART) for identity documents (KYC fraud prevention) maintaining scalability.

Sep 2019 – Sep 2023 · 4 yrs · Delhi, India · Remote

Aug 2019 – Nov 2020 · 1 yr 3 mos · Bengaluru, Karnataka, India

• factorization machines, DeepFM models for CTR optimization
• Candidate Generation and Candidate ranking models for e-commerce advertisers to overcome scaling issues (200M MAU)
• Model deployment at scale inference times
• Lookalikes Modeling using neural networks for cold start problems

Feb 2019 – Nov 2021 · 2 yrs 9 mos · Pune, Maharashtra, India · Hybrid

May 2018 – Aug 2019 · 1 yr 3 mos · Gurgaon, Haryana, India

Mar 2017 – Apr 2018 · 1 yr 1 mo · Gurgaon, India

• Deployed Decision tree model to increase user retention on Zoto, from features ranging from customer demographics to call logs and billing details.
• User segmentation using tf-idf statistics and K-means clustering.
• Random Forest model to target users for a cashback campaign, in order to optimize the marketing spend.
• Cashback optimisation
• Live and Interactive Dashboards on Plotly , Metabase and Redash.
• People Analytics using Mixpanel.

Apr 2016 – Mar 2017 · 11 mos · Noida Area, India

• Product recommender system using collaborative filtering based on cosine similarities for marketplace platform.
• Seller segmentation using clustering Algorithms ( K-means) and dimensional reduction using principal component analysis.
• Hands on experience on Data Visualization, Exploratory Analysis and Dashboards using Tableau Desktop, plotly, Metabase and Redash.
• Responsible for increasing the active seller base on Marketplace platform.
• Web Analytics & Traffic Growth using Google Analytics
• Product Analytics and Intelligence using analytical and Big Data tools(R programming, MySQL, HIVE)
• Hands on experience in Drawing statistical inferences, data mining, regression modeling (Generalized Linear Models), pattern recognition and automation to drive sellers growth in consensus to the company's business.

Jun 2015 – Mar 2016 · 9 mos · Noida Area, India

Jun 2014 – Jul 2014 · 1 mo · Ghāziābād Area, India

• Worked under SYSTEM PRODUCTION DIVISION(SPD) subdivided into :
• 1) Quality Control (QC ). Implemented Statistical Process Control involving application of 7 Quality Tools(Pareto Charts, Fish-bone Diagrams, Control Charts etc.) to control the process's variability as to deliver high end quality products.
• Also Worked on Inventory Management(ABC, VED, FSN etc.) Planning and Forecasting , Inventory Models, Optimizing inventory to minimize Inventory costs
• 2) Production Engineering Group( PEG ) : Integration of SYSTEM(s) of people and equipments.
• 3) Production Planning and Control( PPC ) :Hands on experience on Forecasting as a planning tool,developing forecasting logic and evaluating accuracy of forecasts(MAD,MAPE,MSE)
• POS Data Systems and Basic Strategies for aggregate production planning :
• Level strategy
• Chase Strategy and Mixed Strategy

Dec 2013 – Jan 2014 · 1 mo · New Delhi Area, India

• Worked under the Planning Department at ,One of the leading Design, Engineering and Consultancy organizations of South East Asia.
• Employed Project Management Concept and tool Primavera (Enterprise Project Portfolio Management Software) to design project plans, milestones and evaluate the project's progress and generate MIS reports for several engineering and technical projects.
• Team member, evaluating the Planning Package for ONGC Water Injection North (WIN) platform (located in Mumbai High Field) revamp project.