Open to AI / ML / Full-Stack roles — Melbourne & remote

Vinit Kumar

AI Engineer & Full-Stack Developer

I build production AI systems — from model training to deployed full-stack products.

Melbourne, Australia

vinit.ts

const engineer = {
  role: "AI Engineer & Full-Stack Dev",
  builds: ["RAG", "agents", "fine-tuned LLMs"],
  stack: ["Python", "TypeScript", "Next.js", "FastAPI"],
  deploys: "Vercel", "Railway", // live in production
  status: "open to work",
};

Applied LLM & RAG

Retrieval-augmented systems and agents grounded in real data — multi-source ingestion, vector search, and cost-aware LLM orchestration.

ML & Model Training

Hands-on model work, not just API calls — parameter-efficient fine-tuning (LoRA/PEFT), computer vision, and measurable results.

Production Full-Stack

Shipped software for real users and clients — rules engines, dashboards, and end-to-end apps with testing and deployment.

Featured work

Projects, written up like an engineer

Flagship projects — several deployed live — each written up as a short case study covering the problem, the architecture, and the decisions behind it.

Applied LLM & RAGLive

Vehicle DIY Guide

An agentic RAG system that builds vehicle-specific repair guides from manuals, forums and video — with a live AI mechanic you can ask follow-ups.

source types fused per guide

cost per cached repeat

cost-tiered Claude models

FastAPINext.jsClaudepgvectorRedisPython

Case study Code Live demo

Applied LLM & RAGLive

Aria Voice AI

A human-like AI voice assistant you talk to in the browser — real-time speech in and out, answered from a business's knowledge base via RAG. Paste a URL or PDF to talk to your own.

live APIs in the spoken round-trip

~1s

first response, streamed

~10s

URL/PDF → your own voice agent in

FastAPIWebSocketsDeepgram Nova-3Claude Haiku 4.5ElevenLabs Flash v2.5Supabase pgvector

Case study Code Live demo

Production Full-StackLive

Exempt Development Tool

A government capstone: a rules engine that assesses building proposals against NSW planning law and returns a clear, clause-referenced decision.

25,305

property records searched

Albury

real government client

engineer team

ReactNode.jsExpressGoogle Maps APIJestTailwindCSS

Case study Code Live demo

Production Full-StackLive

MorningByte

A subscription product that emails a short, AI-summarised briefing on your chosen topics every morning — with double opt-in and one-click unsubscribe.

topic categories (multi-select)

Haiku

AI-summarised daily

1×

LLM call per topic, not per reader

FastAPIPostgresClaude HaikuResendRSSRailway

Case study Code Live demo

More work

LoRA Doctor

Fine-tuned a language model on medical Q&A using LoRA — training 0.23% of the parameters of a full fine-tune.

PyTorchPEFTHuggingFaceLoRA

View code

RAG PDF Chat

Chat with any PDF — fully local, no API keys, answers grounded strictly in the document.

FAISSsentence-transformersOllamaStreamlit

View code

GitHub Buddy

A developer tool to configure Git, auto-generate READMEs and manage repositories.

Next.jsTypeScriptPrismaDocker

View code

Toolbox

Skills

Languages

PythonTypeScriptJavaScriptSQLJava

AI / ML

RAGLoRA / PEFT fine-tuningLangChainHuggingFacePyTorchTensorFlowOpenCVClaude API

Web & Backend

Next.jsReactFastAPINode.jsExpressTailwind CSS

Data & Cloud

PostgreSQLpgvectorSupabaseDockerVercelRailwayPrisma

Track record

Experience

IT Support & Operations Lead · Aaditya Rentals

2023 — Present

Melbourne

▸Lead a team of 4, overseeing daily operations and front-line delivery.
▸Run QA and structured defect reporting for the company website and internal systems.
▸Built internal automation integrating the Xero and Zoho Mail APIs for invoicing and debtor workflows.

AI / Full-Stack Developer — Industry Capstone · Albury City Council

2025

Team of 5

▸Built the rules engine assessing building proposals against NSW SEPP planning legislation.
▸Developed the React frontend + Google Maps integration over a 25,305-record property database.
▸Led testing with a Jest unit + integration suite against client acceptance criteria.

Education

Master of Information Technology — Artificial Intelligence

La Trobe University, Melbourne · 2024 — 2026 · GPA 6.0 / 7.0

Bachelor of Computer Science — Data Science

Lovely Professional University, India · 2019 — 2023

Certifications

▸Supervised Machine Learning: Regression & Classification — DeepLearning.AI & Stanford (Coursera)
▸Google Data Analytics Professional Certificate — Google (Coursera)