cv

Research Assistant

Knowledge Systems & IRSL Lab, Stony Brook University (SBU)

Research Assistant focused on autonomous navigation and active kinesthetic learning.

• Minimized human demonstration burden in active kinesthetic learning by integrating PAC learning and bandit algorithms.
• Advancing autonomous navigation for mobile manipulators by developing targeted algorithms on a Segway platform.
• Establishing a scalable humanoid navigation system by engineering a camera and LiDAR calibration setup for the Unitree G1 bipedal robot.

Technical Lead Engineer

Saama Technologies Inc

Technical Lead focusing on generative AI for clinical data.

• Achieved 50% effort reduction in SDTM generation through an NLP-driven ETL pipeline.
• Generated analytical TLF data using LLM mapping with Bedrock, Azure OpenAI and LlamaIndex.
• Adapted Llama2 and Mistral models to the clinical domain via sequential fine-tuning.

Lead ML/ Deep Learning/ Computer Vision Engineer

DisplaySweet

Lead Machine Learning and Vision Engineer.

• Enhanced spatial analysis by developing deep learning algorithms for 3D mesh optimization and floor plan parsing.
• Increased productivity by developing an AI agent for customer analysis.

Deep Learning Software Engineer

Kinara, Inc - MLAI System

Deep Learning Engineer specializing in edge computing optimization.

• Secured a model compression patent by designing an automated mixed-bit quantization and pruning algorithm.
• Optimized edge computing performance by resolving integration issues with video-based segmentation.

Project Officer

Indian Institute of Technology, Kharagpur

Accelerated CNN inference natively on embedded devices by designing a self-decomposing CNN.

Junior Research Fellow (MHRD-CSE)

Indian Institute of Technology, Kharagpur

Achieved autonomous tracking on moving objects using optical flow-based state estimation.

• Validated autonomous navigation systems natively within a Gazebo simulation.

Ph.D Student

SUNY Stony Brook University

Computer Science

Master of Science (Research)

Indian Institute of Technology, Kharagpur

Computer Science and Engineering

• Developing Reusable, Speedy, and Compact Deep CNN Models

B.Tech

Indian Institute of Technology, Kharagpur

Electronics and Electrical Communication Engineering

• Navigational Signal Processing

Computer Science Chairman Fellowship Award

Stony Brook University

Fellowship for Ph.D studies at SBU.

Junior Research Fellowship (MHRD)

IIT Kharagpur

Fellowship for research work.

Best Team Cooperation Award (IARC)

International Aerial Robotics Competition, Beijing

Merit cum Means Scholarship

IIT Kharagpur

Scholarship for undergraduate studies.

Method for automatic hybrid quantization of deep artificial neural networks (US Patent App. 17/112,889)

Inventors: W. Qadeer, R. Hameed, S. R. Uppalapati, A. B. Ghanore, and Kasanagottu Sai Ram

Deterministic Continuous Replacement: Fast and Stable Module Replacement in Pretrained Transformers

arXiv preprint

arXiv preprint arXiv:2511.18670

Fitness based layer rank selection algorithm for accelerating CNNs

IEEE ICIP 2019

IEEE International Conference on Image Processing

HSD-CNN: Hierarchically self-decomposing CNN architecture

ICVGIP 2018

ACM, 11th Indian Conference on Computer Vision, Graphics and Image Processing

Nonseparable filters for images in the block DCT domain

IEEE ICIP 2018

2018 25th IEEE International Conference on Image Processing (ICIP)

Video based person re-identification by re-ranking attentive temporal information in deep recurrent convolutional networks

IEEE ICIP 2018

2018 25th IEEE International Conference on Image Processing (ICIP)