Dec 2017 – May 2019 · 1 yr 5 mos · Bangalore

• Developed end to end ML framework for e-commerce product matching
• Integrated multiple machine learning models like text, image and other together
• Used deep learning models like CNN, LSTM, Bi-LSTM CRF for product matching
• Lead the project from requirement gathering to deployment with 2 person team
• Collaborated with multiple teams for product adaptation and deployment
• Bridged the gap between business and technology by presenting the business impact of the project
• Trained other members of the team to grow in Machine learning domain

Apr 2016 – Dec 2017 · 1 yr 8 mos

• Developed scalable document classification engines using XGBoost and Random Forest to categorize documents across 300+ classes, outperforming commercial solutions.
• Built a fully open-source document processing pipeline leveraging R, Tesseract OCR, an in-house classification system, and automated deployment, replacing subscription-based tools and generating multimillion-dollar annual savings.

Dec 2014 – Apr 2016 · 1 yr 4 mos

• Working on classification of mortgage document in 300+ classes and separate document boundaries. Used Random forest algorithms to successfully classify documents with above 85% accuracy and Modified Hidden Markov Model for document separation with above 84% accuracy on very large dataset.

Mar 2014 – Aug 2014 · 5 mos · Bangalore

• Worked on various challenging projects like converting SAS to R, alalyzing EEG data and predicting seizure, conversation mining

Jul 2013 – Mar 2014 · 8 mos · Bangalore

• Joined in video drivers team. Working very closely with the hardware TI produces and knowledge of how video works. Made the code MISRA C compliant. Did major bug fixes in drivers. Added interrupts to solve the issue of overflow

May 2012 – Jul 2012 · 2 mos · Bangalore (India)

• Did a project with buffer management during AV playback. Decoupled the use of 2D tiler and 1D ion buffer for processing and display.