Skip to content
View SyN-droMe's full-sized avatar

Block or report SyN-droMe

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
SyN-droMe/README.md

Ayush Kuriakose Chacko

AI/ML Engineer: computer vision, predictive ML, and agentic workflows. B.Tech CS (VIT Chennai). Based in Bengaluru, open to relocation.

Recently wrapped up an apprenticeship at Mercedes-Benz R&D India, where I worked on computer vision, edge deployment, knowledge distillation, and gesture recognition on Jetson Orin. Open to AI/ML Engineer and Data Scientist roles. Currently contributing to open source.

Projects

  • shrinkwrap — drop-in Python wrapper that cuts LLM API cost via prompt caching, context summarization, and response caching. ~27% lower cost across Anthropic and AWS Bedrock
  • ScholarAI — RAG pipeline (LangChain + ChromaDB) for research-paper Q&A, with dual LLM backends (Gemini + local LLaMA via Ollama) behind a FastAPI + Streamlit interface.
  • Time Series Project — Forecasting heart-rate (BPM) from a live MAX30105 pulse-sensor stream during gameplay, using a pretrained PatchTST transformer — 86–88% accuracy vs. ground-truth moving-average BPM.

Stack

Python · PyTorch · TensorFlow · LangChain · scikit-learn · Docker · AWS · GCP

LinkedIn

Pinned Loading

  1. emulsion emulsion Public

    An interactive map of 3,222 recipes. Route between dishes by what they are made of or how they are made.

    Python 2

  2. bulkhead bulkhead Public

    A tool-calling agent harness that keeps working when tools fail. Retries, timeouts, loop detection, and honest error reporting, benchmarked against injected failures.

    Python 4

  3. shrinkwrap shrinkwrap Public

    Compresses LLM prompts to cut token cost while preserving meaning, with measured before/after numbers rather than claims.

    Python 6

  4. ScholarAI ScholarAI Public

    A RAG pipeline that answers questions from a document corpus with grounded, citable answers.

    Python