Sanjeev Kumar
Software Engineer with 6+ years at Google, LinkedIn, and Amazon, applying ML/AI to e-commerce, finance, and cybersecurity. Developed ranking models at Google, enhanced fraud detection at LinkedIn, and led ML projects at Amazon. M.Tech from IIT Bombay, specializing in deep learning, NLP, and AI-driven solutions.
- Role
AL ML Engineer
- Years of Experience
3.9 years
- Professional Portfolio
View here
Skillsets
- ML
Professional Summary
3.9Years
- Jan, 2026 - Present 6 months
Machine Learning Engineer
Meta - Jan, 2022 - Jun, 20231 yr 5 months
Software Engineer
Google - Aug, 2021 - Jan, 2022 5 months
Senior Applied Research Engineer
LinkedIn - Jul, 2019 - Oct, 2019 3 months
Senior Data Scientist
JOTTER.AI - Jan, 2020 - Aug, 20211 yr 7 months
Applied Scientist
Amazon
Applications & Tools Known
Machine Learning
Data Science
Natural Language Processing
Computer Vision
Artificial Intelligence
Work History
3.9Years
Machine Learning Engineer
MetaJan, 2026 - Present 6 months
Software Engineer
GoogleJan, 2022 - Jun, 20231 yr 5 months
Developed various filtering/ranking based strategies to improve landing page quality for user by leveraging clicks+relevance scores.
Developed query transformation techniques following relevance feedback paradigm to create more user engagement on SRP results.
8% relevance gains with 3% coverage drop was incorporated on SRP relevance reporting pipeline
Senior Applied Research Engineer
LinkedInAug, 2021 - Jan, 2022 5 months
Applied Scientist
AmazonJan, 2020 - Aug, 20211 yr 7 months
Demand for newly launched mobile devices is harder to predict because of lack of data.
Developed machine learning models based on transfer learning, percentile features and attention based neural network models to get 40% mean absolute error on newly launched mobile phones.
Senior Data Scientist
JOTTER.AIJul, 2019 - Oct, 2019 3 months
Achievements
- Granted State Board Merit Certificate for being in top-100 students in 10th and 12th examination
- Received honorarium in IIT Bombay to develop IR pipeline for Anti Terrorism Squad, Mumbai
- Awarded 10/10 pointer in R&D Machine Learning project and Information retrieval in IIT Bombay
Major Projects
9Projects
Landing Page Quality Improvement for Shopping Platforms
Jan, 2022 - Jan, 20231 yr
Developed various filtering/ranking based strategies to improve landing page quality for user by leveraging clicks+relevance scores. Developed query transformation techniques following relevance feedback paradigm to create more user engage- ment on SRP results. 8% relevance gains with 3% coverage drop was incorporated on SRP relevance reporting pipeline.
Feed Quality Improvement using Classification Models to detect Money Fraud on Platform
Aug, 2021 - Jan, 2022 5 months
Developed text-based feature extraction pipelines to gather signals pertaining to feelings, money, and bank/contact information on feed post. Prepared text pre-processing pipelines to train FastText based text embeddings. 8% recall gains and 4% precision gains were added to an upgrade in the money fraud detection pipeline.
Attention Based Demand Prediction Model for newly launched wireless devices
Aug, 2020 - Aug, 20211 yr
Demand for newly launched mobile devices is harder to predict because of lack of data. Developed machine learning models based on transfer learning, percentile features and attention based neural network models to get 40% mean absolute error on newly launched mobile phones.
Multilabel Text Fitment prediction Model for automobile products
Jul, 2020 - Aug, 2020 1 month
Developed a pipeline to train a multi-label neural network classifier to predict FITMENT for each automobile product with 97% precision and 73% recall.
Duplicate Suppression in autocomplete search suggestions based on user behaviour
Apr, 2020 - Jul, 2020 3 months
Autocomplete system is great tool to promote guided shopping by educating customer about the available products in catalogue. Different ways to articulate the customers intent bring duplicates in search suggestions. Developed framework based on customers interaction with catalogue & MMR(Maximal marginal relevance) to suppress duplicate suggestions, prepare query vector & rank suggestions based on relevance-diversity tradeoff.
Device Fingerprinting with zero permissions for Fraud detection and User identification
Jan, 2020 - Apr, 2020 3 months
Developed a framework for user identification on platform to re-identify old users and discriminate new users based on users personal configuarion. Acheived around 99.99% precision and 99.96% recall for known devices after collection of 50,830 device finger- prints over time from 2,239 devices.
Image based Virtual Try-On for Fashion Models Wearing Custom Outfits
Jun, 2019 - Dec, 2019 6 months
In e-commerce fashion industry, there is need to generate high-resolution images of fashion models wearing a custom outfit under an input body pose. Developed Generative Adversarial Networks to generate fake human bodies for e-commerce fashion catalog.
Vision based structural information extraction for domain specific knowledge graphs
Jul, 2016 - Jun, 20192 yr 11 months
Tables extraction from e-commerce html pages based on vision based page segmentation and following SEAL pat- terns to expand the size of the set Single program logic based on vision extraction rules successfully worked with ecommerce websites like Ebay, Alibaba, Amazon, Flipkart, Groupon, Goodreads, Makemytrip, Snapdeal, Shopclues, Paytm, Mercari and Coursera
Rollcheck calculations on the platform along with Forex trades reconcilation
Jul, 2015 - Jun, 2016 11 months
Developed automated tools to verify the consistency of automated transactions that took place on platform including Forex trades based on various design patterns
Education
MTech-3 Year RA
IIT Bombay (2019)Graduation
HPU UIIT Shimla (2015)Intermediate/+2
HPBSE GSSS Takka (2010)Matriculation
HPBSE JBP Nari (2008)