Feb 2022 – Apr 2024 · 2 yrs 2 mos · Bengaluru, Karnataka, India · On-site

• Building ML platform to serve personalized recommendations to Amazon Sellers on Amazon's largest Seller Facing platforms across all Amazon marketplaces.
• Tech Stack - Spark, Python, DynamoDB, AWS

Apr 2021 – Jan 2022 · 9 mos

• Responsibilities include designing Machine Learning solution, automating model training, deployment by building scalable APIs and dash-boarding for performance monitoring by working closely with product team to understand business requirements. Responsible for end to end project management and mentoring a small team.
• Building Flights Fare Prediction model which powers the new FareLock product offered on MMT and GI. Used GBM and quantile regression to predict flight fare movement. Deployed the system using an event driven deployment using Kafka, Aerospike, ElasticSearch.
• Built dynamic pricing system for real-time pricing of Hotels funnel. Experimented with several algorithms: Multi-Armed bandit, DNN Contextual Bandits and Reinforcement Learning. Used airflow to automate training pipelines which fetch data from multiple data sources (Athena, S3, Redshift) using different tools (Python, PySpark, SQL) to create more than 100 different static and real-time features.
• Developed and deployed an entity embeddings based NN model for Indian Railways Waitlisted PNR confirmation prediction problem. Model performance was better than the leading similar probability providers.
• Helped improve the event level data logging used at organisation level for building ML models and monitoring performance by debugging the data extensively and helping build anomaly detection tools.
• Presented work at external conferences & meetups and internal tech sessions. Actively involved in hiring process of prospective candidates.
• Tech Stack - Python, SQL, Redash, Airflow, Redis, AWS, PySpark, Pandas, Kafka, Aerospike

May 2018 – Apr 2021 · 2 yrs 11 mos

Dec 2015 – Apr 2018 · 2 yrs 4 mos · Gurgaon, India

• Helped setup the entire data infrastructure for the organisation from scratch including event logging, visualisation and dash boarding tools, ML model deployment framework, data pipelines etc.
• Used machine learning, natural language processing, image processing to make app experience personalised for each user and improve the efficiency of internal teams by automating repetitive tasks.
• Conducted statistical analysis and deep dived large amounts of data using advanced analytics tool to identify actionable insights to help internal teams (product, marketing, medical affairs, BD) take strategic decisions

Jul 2013 – Dec 2015 · 2 yrs 5 mos · Gurgaon, India

• Involved in end to end delivery of multiple short and long term projects which required using statistical analysis and machine learning techniques such as multi-variate regression analysis, K-means clustering, greedy-algorithm based optimization, test-control analysis, ROI evaluation, scenario simulation etc. These projects were primarily executed using SAS, SQL, R and Excel.
• Projects undertaken includes Marketing Mix Optimization, Promotion Response Modeling, Sales Force Sizing, Test & Control Analysis

May 2011 – Jul 2011 · 2 mos · Coventry, United Kingdom

• Developed a methodology to improve robustness of a complex product comprising of several components by minimization of interaction between the components.
• Also developed an algorithm to intelligently generate explicit liaison charts of complex systems.