Suryaprakash Senthil Kumar

My Bio

I’m an AI Engineer based in San Francisco, USA, focused on building scalable backend systems and deploying AI models for SaaS products that boost productivity. I love bringing research ideas into production and consistently strive to bridge that gap. Outside of work, I enjoy attending hackathons, going on hikes, playing volleyball, and listening to podcasts 😊

Currently...

Hackathons

Attending a hackathon in SF (Oct 11/12 2025) and I am actively looking for teammates 🕵!!
If you are interested let’s connect and build something awesome together!
AI - Research

Exploring model context compression techniques, researching how to optimize long-context reasoning and improve LLM efficiency through better retrieval and pruning strategies.

Recents

Reflections on my recent work @Proxis

September, 2025
Cursor-v2: a coding assistant focused on frontend development

June, 2025
GPU Chess Hackathon @Strong Compute

September, 2024

NOTE: Portfolio design was inspired from codewithsadee

Resume

Education

Georgia Institute of Technology
2022 — 2024
Master of Science in Robotics (Artificial Intelligence & Perception)
GPA: 4.0/4.0

Relevant Coursework:

Machine Learning | Algorithms | Natural Language Processing | Computer Vision | Deep Learning | Operating Systems | Reinforcement Learning | High Performance Computing
SSN College of Engineering (Anna University)
2018 — 2022
Bachelors in Mechanical Engineering
GPA: 3.7/4.0

Relevant Coursework:

Probability | Statistics & Numerical Methods | Machine Vision

Experience

AI Software Developer

Proxis

San Francisco, CA

Python TypeScript FastAPI Redis Supabase AWS asyncio OpenAI

Duration: May 2025 — Present
1. Architected a RAG pipeline with semantic search capabilities; implemented a hybrid scoring schema, improving context recall by 25% through analyzing relevancy and recency of knowledge base
2. Streamlined the agentic workflow by minimizing tool calls by 50%; constructed an eval-agent with GPT-4o as a judge, enabling feedback-based self-correction of email drafts
AI Engineer

Chima

San Francisco, CA

Python OpenAI

Duration: Apr 2025 — May 2025
1. Worked on agentic AI systems
Machine Learning - Research Affiliate

BrainML Laboratory (Georgia Tech)

Atlanta, GA

Python PyTorch Transformers Distributed Computing Graph Neural Networks

Duration: Aug 2024 — Apr 2025
1. Modeled spatio-temporal graphs to capture dynamics in a multi-agent mice environment, improving motion prediction accuracy from 68% to 80% using transformers
Computer Vision and Deep Learning Intern

Swaayatt Robots Private Limited

Bhopal, India

Python PyTorch

Duration: Oct 2021 — May 2022
1. Researched model pruning techniques for object detection and tracking on image data from an L4 autonomous vehicle