Achintya K J

Hi, I'm |

Graduating 2027 · Open to full-time roles in AI research, LLM security, and applied ML engineering.

AKJ

01 — About

Most of my work sits where machine learning meets security. Right now that means building LLM red-team suites against the OWASP LLM Top 10, and an eBPF tool that watches the kernel to find Shadow AI usage inside client networks: observability at the OS layer rather than the API layer. Before that, I trained graph neural networks to spot social-engineering patterns in conversation data. I care most about systems that stay interpretable when they are under adversarial pressure.

02 — Work Experience

Pragya Cyber

AI Pentesting Research & Development Intern

Aug 2025 - present

Building SecuraGPT, a multi-cloud B2B platform for securing AI agents through automated red teaming, observability, and vulnerability assessment. Involved end-to-end in the research, design, and implementation — including a custom policy-enforcement engine and observability stack for centralized telemetry, compliance monitoring, and security-policy evaluation. Alongside that I research and engineer test suites against the OWASP Top 10 for LLMs (prompt injection, insecure output handling), and a kernel-level eBPF monitor (via Grafana Beyla) that detects and quantifies Shadow AI usage on client networks.

PwC India

Launchpad Program Trainee

Feb 2026 - Jul 2026

Selected for the PwC Launchpad program, a structured consulting and digital upskilling initiative. Gained hands-on experience in business consulting frameworks, corporate problem-solving, and technology solutions under the mentorship of PwC consulting leaders.

E2E Management Services

Summer Intern

Jun 2025 - Jul 2025

Completed a structured deep-dive across ML, Deep Learning, and Reinforcement Learning, applying concepts through hands-on project work. Engineered an AI Legal Chatbot capable of answering user queries grounded in uploaded legal documents, fine-tuned on a corpus of Indian legal cases for domain accuracy.

03 — Projects

Selected work

Applied ML, security tooling and product experiments. Click a project to watch its demo.

Demo

Source

Download

Toki

2026

A screen-time tracker for Windows and Linux (Hyprland) that keeps everything in a SQLite file on your own disk.

Rust

windows-rs

Linux (Hyprland)

SQLite

Tauri v2

SvelteKit

Demo

Source

Luma AI Interview

2025

An AI hiring pipeline: resume scoring, a proctored coding test, then a voice interview with an AI agent.

Next.js

FastAPI

Convex

LiveKit

Gemini

MediaPipe

Demo

Source

PaperForge

2026

Turns arXiv papers into implementation prompts for coding agents.

Python

FastAPI

React 19

Tailwind

PyMuPDF

04 — Open Source

OSS contributions

Pull requests I've opened on other people's repos.

05 — Achievements

Building under pressure

I have competed in 11 hackathons and CTFs, from campus sprints to national stages. A few standout placements below; the rest were just as valuable for the people and the problem sets.

Feb 2026Challenge Author