Shreyas Sinha
Goa
Overview
1
1
year of professional experience
Work History
Upcoming Software Development Engineer Intern
Amazon
01.2026 - Current
Machine Learning Intern
AGC Asia Pacific
05.2025 - 07.2025
- 3D LiDAR Object Detection & Tracking: Developed a deep learning pipeline for 3D object detection and real-time multi-object tracking on LiDAR point clouds. Adapted DeepSORT for 3D, integrating custom backbone architectures with Kalman filtering for trajectory association. Increased inference speed by 20% and reduced missed detections by 18%.
- Graph Neural Network for 3D Shape Classification: Designed and implemented a lightweight GNN for 3D shape classification, achieving 89% accuracy on ModelNet. Reduced training time and computational requirements, demonstrating efficiency over conventional baselines.
Undergraduate Researcher
TCS Datalabs
01.2025 - 07.2025
- LLM Research & Multi-Modal Reasoning: Engineered RAG pipelines for LLMs, architecting multi-modal (text, figure, table) document ranking for 300+ scientific QA tasks. Achieved a 62% lift in MRR, from 0.50 to 0.81, outperforming top open-source models.
- Scientific QA Impact: Raised answer accuracy from 50% to 70% and Recall@3 from 69% to 85% on M3SciQA, SPIQA, DocGenome via prompt-driven retrieval, per-paper and pairwise ranking, and scalable table/figure parsing.
Education
B.E. (Hons.) - Electronics and Communication Engineering, Data Science
Birla Institute of Technology and Science
11.2025
Skills
- Python
- C
- Java
- SQL
- Linux Shell
- PyTorch
- TensorFlow
- FastAPI
- Flask
- MLflow
- Machine Learning
- Deep Learning
- Generative AI
- LLMs
- NLP
- CV
- Networking
- Time Series
- Jupyter
- Docker
- Linux
- GitHub
Timeline
Upcoming Software Development Engineer Intern
Amazon
01.2026 - Current
Machine Learning Intern
AGC Asia Pacific
05.2025 - 07.2025
Undergraduate Researcher
TCS Datalabs
01.2025 - 07.2025
B.E. (Hons.) - Electronics and Communication Engineering, Data Science
Birla Institute of Technology and Science