Jul 2023 – May 2025 · 1 yr 10 mos

• Train large offline multi-modal models (Lidar + camera sensor fusion) for auto labeling
• Deploy latency critical 3D obstacle detection models by leveraging TensorRT and custom CUDA kernels
• Continuous data mining / model training using PyTorch, kubeflow and airflow.

Dec 2021 – Jul 2023 · 1 yr 7 mos

• Working on LiDAR based 3D obstacle detection problem.
• Responsible for end to end pipeline from data mining, model training to model deployment on trucks.

Jul 2021 – Dec 2021 · 5 mos

Sep 2019 – Mar 2021 · 1 yr 6 mos

• Teaching:
• CSE 150B - Intro to AI: Search and Reasoning
• CSE 251B - ML: Neural Networks and Pattern Recognition
• Research:
• Finding Rectilinear Steiner Minimum Tree using DeepRL: Using deep Reinforcement Learning to tackle the NP-complete combinatorial optimization problem of finding the RSMT for input pointset. Guide: Prof. Lawrence Saul
• Latent Domain Discovery For Domain Adaptation: Formulated domain adaptation as discovering latent source domains and aligning conditionals using novel probabilistic extension to MMD. Guide: Prof. Manmohan Chandraker

Jul 2016 – Jul 2019 · 3 yrs · Mumbai Metropolitan Region

• i) Using Statistics and Machine Learning models to gain insights in the market data so as to make accurate future price predictions and make profitable trades.
• ii) Successfully created fully automated trading strategies keeping risk appetite in mind.