Open to AI/ML & SWE internships

Abhishek Bevinkatti


FINAL-YEAR B.E. CS (AI & ML) · JAIN COLLEGE OF ENGINEERING, BELAGAVI

Building systems that are useful & measurable.

Retrieval pipelines, multi-agent orchestration, and the evaluation harnesses that keep them honest. A few of the systems I've shipped — from a 0.927 RAGAS research assistant to a browser-native PDF editor that never touches a server.

View work Say hello →
0.927 RAGAS SCORE· 38% LATENCY REDUCTION· <15ms P99 RETRIEVAL· 90%+ FRAUD DETECTION ACCURACY· 12 PDF TOOLS, ZERO SERVERS·
About

Somewhere between a CS degree and a habit of building things.

The non-resume version.

I'm a final-year CS (AI/ML) student who somehow ended up running a handful of real, deployed systems off free-tier infrastructure and pure stubbornness. It started as "something concrete for the resume" and turned into an actual habit — I'm more likely to spend a weekend building an open-source project than studying for an exam.

Most of what I build starts as a half-baked idea that eats my weekdays and evolves into a solid project. I lean hard on measuring things — RAGAS scores, latency numbers, false-positive rates — because quality should be something you can show, not just claim.

Outside of that: way too many side-project ideas lying around in my notes.

Off the clock
  • Based in Belagavi, Karnataka, India
  • Finding ways to make things work within tight constraints, especially budget ones
  • Strong, unsolicited opinions about font embedding
  • Will not ship a number I haven't measured myself
  • Prefers music in the background while building and coding
  • Stuck choosing which to start next: Elden Ring or The Last of Us.
Selected Work

The Solid recent ones, all held to the same standard —
does it work, and can you prove it ?

Every project ships with the numbers attached.
Nothing here is a demo you can't measure.

Multi-Agent

Multi-Agent AI Research Assistant

A RAG platform that scores its own answers before it gives them to you.

LangGraph orchestrates parallel research agents across a FAISS retrieval layer, with self-correcting evaluation loops and per-response RAGAS scoring built in — not bolted on. Real-time SSE streaming and structured latency, quality, and error tracking, deployed live on Hugging Face Spaces.

LangGraphGroq · openai/gpt-oss-120bFAISSFastAPIDocker
Open → GitHub →
0
RAGAS Score
0
Faithfulness
0% ↓
Latency (6.5s → 4.0s)
<0ms
P99 Retrieval
Open Source

PDFZero

A 12+ tools PDF editor that never sends your file to a server.

Text editing, merge, split, compress, redact, unlock, OCR and more — 100% client-side. A font-classification engine maps 40+ font families for pixel-accurate text edits, with offline OCR via Tesseract.js and a code-split bundle under 42KB on first load.

ReactVitePDF.jspdf-libTesseract.jsZustand
Open → GitHub →
0
PDF Tools
0%
Client-side
0+
Fonts Mapped
<0KB
Initial Payload
Core Complete

DevPulse

Webhook delivery that survives crashes, 5xx errors, and duplicate sends.

A self-hosted reliable webhook delivery system. Events are committed to PostgreSQL with a transactional outbox, published through Redpanda, and delivered by a worker as HMAC-signed requests, with durable retries, idempotent ingestion, tenant isolation, and SSRF-safe destination validation. Verified locally with a deterministic receiver that fails once, then recovers.

NestJSPostgreSQLRedpandaNext.jsGraphQLTypeScriptDocker
GitHub → Architecture →
0
Tests Passing
503→204
Retry Path Verified E2E
HMAC
SHA-256 Signed Deliveries
0txn
Event + Outbox Write
At-least-once delivery. Receivers dedupe via Idempotency-Key.
Open Source

Loop Watchdog

A kill switch for AI agents that won't stop retrying.

A local-first FastAPI proxy sitting in front of OpenAI-compatible coding agents, scoring traffic across five signals — retries, file churn, error signatures, edit-oscillation, request growth — and blocking runaway loops before tokens burn. A Cloudflare Worker control plane pushes HMAC-signed alerts to Slack and email.

PythonFastAPITypeScriptCloudflare WorkersD1pytest
GitHub →
0
Detection Signals
0+
REST Endpoints
Blocks on HTTP 409 before the next retry fires.
Open Source

RAG Evaluation CLI

Benchmark any RAG system without hitting an external API.

A framework-agnostic CLI for deterministic RAG evaluation — schema auto-detection for JSON/JSONL/CSV, RAGAS-style metrics, and multi-system comparison for faster iteration.

PythonCLIRAGAS
GitHub →
0k+
PyPI Downloads
0
Evaluation Metrics
Complete

StyleSync Elite

Fashion generation tuned to run on a GPU you can actually afford.

Stable Diffusion XL Turbo optimized for low-latency inference on consumer-grade hardware, with a FastAPI backend and a prompt-controlled generation pipeline for real-time fashion outputs.

Stable Diffusion XLPyTorchFastAPI
GitHub →
5–7s
Inference Time
0GB
VRAM Required
Complete

Fake Job Postings Detector

Flags a scam listing before you waste an application on it.

TF-IDF feature extraction and ML classification identify suspicious postings with explainable, confidence-scored risk insights — wrapped in a Streamlit interface built for fast, real-time checks.

Scikit-learnNLPStreamlit
GitHub →
~0%+
Detection Accuracy
0% ↓
False Positives
Open Source

Shipping libraries people actually install.

Beyond personal projects — contributing upstream,
publishing packages, and building in public.

Community

Open Source & Community

Building tools developers actually use.

rag-harness is a framework-agnostic RAG evaluation toolkit published on PyPI with 2.5k+ downloads, built around deterministic evaluation and reproducible benchmarking. Alongside it, I've published multiple Python packages reaching 7.5k+ cumulative PyPI downloads, focused on developer tooling and practical AI infrastructure.

I actively contribute to open-source AI tooling through merged pull requests, publish libraries, and enjoy improving existing developer tools just as much as building new ones.

7.5K+
Total PyPI Downloads
2.5K+
rag-harness Downloads
4
Published Python Packages
Active
Open Source Contributor
The System

Stack — the layers things actually run on.

Agents & Retrieval
LangGraph LangChain Multi-agent systems FAISS RAGAS Sentence-Transformers
Backend & Infra
FastAPI NestJS REST APIs GraphQL SSE streaming Docker
Data & Messaging
PostgreSQL Kafka / Redpanda
Frontend
React Next.js Vite PDF.js
Core Languages
Python TypeScript JavaScript Java C / C++
Timeline

Recent internships, then the degree behind them.

Experience
AI & ML Intern
Edunet Foundation × IBM SkillsBuild — Remote
DEC 2025 — JAN 2026
Cybersecurity with GenAI Intern
Vodafone VOIS × Edunet Foundation — Remote
DEC 2025
Education
B.E. Computer Science (AI & ML)
Jain College of Engineering, Belagavi
EXPECTED JUN 2027
Get in Touch

Seeking AI/ML & SWE roles.

If something here caught your attention, I'd love to hear from you. Whether it's AI, software engineering, open source, or just a quick chat.

Reach-out EMAIL → projectsGITHUB → Connect LINKEDIN →