Prathamesh Anvekar

I'm a backend & ai developer based in Pune, India. I was a computer nerd as a child and I'm a computer nerd as an adult. I build software that I'll love to use.

Currently learning file servers and CDNs with AWS S3 and CloudFront, in Go.

Recently scored 6/6 in FamPay CTF 2026.

Projects

Ergen

Go GitHub

A fast Go CLI that resolves enclosing functions via AST analysis to generate context-aware error-handling. It handles custom type resolution and project configs dynamically, completing end-to-end in 6.75ms. Validated against a 50-function corpus with AST-corruption safety.

Kernel Sandbox Engine
C GitHub

A user-space sandboxing engine that locks down processes using seccomp-BPF filters. Running in a secure fork/exec pipeline, it restricts file, network, and process access, with auditing to safely run untrusted code in CI/CD.

Aegis
FastAPIPython GitHub

A security gateway for LLMs built as a PII and prompt injection filter. It uses regex and classifiers to protect RAG pipelines, backed by a local quantized Gemma 3 instance running with LoRA adapters.

Open Source

Work Experience

Qriocity, Machine Learning Intern

Three months shipping machine learning that had to hold up in production: a custom U-Net that lifted brain-tumour segmentation from 89 → 98%, a HoVerNet + GNN pipeline reading scans in under 2s, and a shelf of other models across imaging and time-series.

Achievements

Top 150 of 80,000+ in Amazon ML Challenge 2025.

Perfect 6/6 in FamPay CTF 2026.

Toolbox

I think in Go, C, and Python, write JavaScript when I have to, serve with FastAPI and Node.js, store in PostgreSQL and Turso, ship with Docker and AWS S3, and train models with PyTorch.

External Links