Aug 2022 – May 2025 · 2 yrs 9 mos · Bengaluru, Karnataka, India · Hybrid

• Worked with Amazon Ads team to use MultiModal Distributed Machine Learning for enforcement of Advertisement moderation policies across images, video and text.

May 2022 – Aug 2022 · 3 mos

• Worked on algorithms for vehicular sensor data perception + aggregation to generate fresh and accurate maps.

Apr 2018 – Apr 2022 · 4 yrs

• Worked on street level imagery problems.
• Prime focus involved 3D reconstruction and localization of street-level features (like signboards, shops, etc) solely from cellphone recorded video data using computer vision, deep learning & computational geometry. Have filed 2 US Patents in this domain.
• Idea here is to further the goal towards a "Self-Healing" Map with minimal human assistance.

Jul 2017 – Mar 2018 · 8 mos

• Worked in the LENS team (Core Maps Engineering) to develop and engineer Deep Learning & Computer Vision algorithms to aid digital mapping worldwide.
• Devised advanced machine vision algorithms for the automatic segmentation and extraction of building rooftop polygons directly from satellite imagery for the 2D Building Footprint Identification (2DFP) project currently in production.

Jul 2020 – Aug 2022 · 2 yrs 1 mo · Bengaluru, Karnataka, India

• Worked on -
• Neural 3D reconstruction.
• Depth + image fusion in CUDA and C++.
• KD-tree building based on ray-tracing.

May 2016 – May 2017 · 1 yr · Lemont, Illinois

• I served as a Research Aide at the Argonne Leadership Computing Facility.
• Applied Deep Learning & Parallel Computing in High Energy Particle Physics using Convolutional Neural Nets
• Worked to scale Distributed Deep Learning models across a 126 node NVIDIA Tesla K80 GPU cluster (Cooley) & a 3,624 node Intel Xeon Phi Knights Landing (KNL) (60+ cores per node) cluster (Theta) to identify particles resulting from collision events at the Large Hadron Collider (LHC), CERN, Geneva, Switzerland detected at the ATLAS detector.

Mar 2016 – Sep 2016 · 6 mos · Chicago

• I've held the position of a Research Assistant in the field of Machine Learning at IIT Stuart School of Business funded by Intel Grant under Prof. Matthew Dixon.
• My work was mainly in the field of high frequency algorithmic trading to develop machine learning algorithms & parallel Python & C++ libraries for trade recommender engines to maximize profit attainable from the trade of various US Treasury Instruments & Futures.
• C++ parallel libraries are designed to run on Intel Xeon Phi Coprocessors (Knight"s Corner)

Jun 2014 – Jul 2014 · 1 mo · Defence Institute of Advanced Technology (D.I.A.T), Pune, India

• Studied hostile movement patterns of maritime vessels and behavioral analysis methods.
• Developed project titled Incremental Self Learning Hostile Object Tracking System using Artificial Neural Networks which calculates a probability of hostility of each entity by performing automated movement behavioral analysis.
• Project implemented using MonoGame® Framework in C# for front-end and MATLAB for back-end ANN training.

Jun 2013 – Jul 2013 · 1 mo · Gurugram, Haryana, India

• Project involved upgradation work of the Issue Tracking Website developed in ASP.NET C#
• Upgradation involved porting from VB.NET to C#.NET, UI Redesigning, and back-end code optimizations.