I specialize in efficient AI systems, LLM optimization, quantization, and computer vision. I am graduating in May 2026 and actively seeking AI/ML Engineer and Applied Research roles.

GitHub LinkedIn Resume

🚀 Key Projects

Distribution-Aware Companding Quantization (DACQ)

Developed a lightweight post-training quantization framework for LLMs that models layer-wise weight distributions and applies non-uniform companding for efficient bit allocation. Integrated activation-aware scaling to preserve downstream accuracy on models such as LLaMA and Qwen.

Stack: PyTorch • HuggingFace • CUDA • Statistical modeling


Efficient Data Pipelines for Vision–Language Models

Clustered prompt embeddings for hybrid autoregressive transformers to reuse lower-scale generated images and reduce redundant computation. Improved computational efficiency through embedding clustering and distributed experimentation.

Stack: PyTorch • Ray • HPC (Slurm)


Core-Set Selection for Incremental Learning

Analyzed dataset characteristics to construct compact core-sets that preserve downstream model performance during incremental updates, improving memory efficiency without significant accuracy degradation.

Stack: PyTorch • Data analysis • Model evaluation


💼 Work Experience

Senior Research Engineer — Toshiba Software India (R&D)

2019 – 2024

Led applied AI research projects in industrial computer vision and manufacturing systems.


🎓 Education

New York University (NYU) Tandon School of Engineering

M.S. Electrical Engineering — Expected May 2026
Coursework: Machine Learning, Deep Learning, Computer Vision, NLP, Efficient AI, Probability & Statistics


Indian Institute of Technology (IIT) Tirupati

B.Tech. Electrical Engineering — 2020


🛠 Technical Skills

Languages: Python, C/C++, SQL
Frameworks: PyTorch, TensorFlow, HuggingFace, Ray, MLflow
Systems: CUDA, Distributed Training, Docker, Linux, Slurm
Focus Areas: LLM Optimization, Quantization, Efficient AI Systems, Computer Vision