Krish
Shroff

• ›Developed internal trading and monitoring tools for co-located exchange platforms, contributing to latency-sensitive infrastructure used in global electronic trading.
• ›Contributed to the design of an internal packet capture tool for analysing network traffic and reporting latency metrics across high-speed trading systems.
• ›Researched AI edge inference on FPGAs for real-time decision support in latency-constrained trading environments.

• ›Led development of RentaCheck, a CNN-based damage detection platform for construction machinery, managing an 8-member team with ML efforts focused on YOLOv8.
• ›Built scalable infrastructure using AWS S3, EC2, and SageMaker for image storage, real-time inference, and automated weekly model updates.

• ›Built an AI risk management system for commercial insurance underwriters — 20% improvement in risk assessment accuracy using HuggingFace Transformers and LLaMA 3.1.
• ›Analysed image data from 10,000+ commercial spaces; deployed models via AWS SageMaker with RAG for policy analysis, increasing report speed by 45%.

• ›Processed and analysed stock data to generate actionable insights from historical trading patterns.
• ›Collaborated on publication tasks including streamlining BibTeX entries for a EuroVis short paper.

Languages

Frameworks

ML / AI

Infrastructure

Libraries