Adithiyan Rajan_

AI Engineer & Builder

I build LLM systems, agentic workflows, and AI tools for engineering teams.CS grad student at uOttawa. Open to collaboration on interesting AI problems.

hot takes on AI1 / 35

About me

Hi, I am Adithiyan, an AI engineer based in Ottawa. I build LLM systems, agentic workflows, and developer tools that turn ideas into impact.

Currently focused on: making language models reliable and useful in production, designing agent architectures, and creating tools that help engineering teams ship faster.

// TODO: understand the question

education

Master of Computer Science · University of Ottawa
B.Tech, Computer Science Engineering · Vellore Institute of Technology

currently

BuildingAgentic workflows and LLM tools for engineering teams
ReadingPapers on evals, agent reliability, and long-context reasoning
ExploringMulti-agent orchestration patterns and structured outputs
Side questsOttawa Youth Council, Toastmasters, side projects

Where I've shipped things

The roles that turned into actual impact.

01

Ericsson

AI/ML Intern · Ottawa, Canada

May 2025 to Dec 2025

Built an LLM log summarizer that saved engineers 10 min per bug across 100+ incidents a week

02

Vector Institute for AI

Machine Learning Associate · Toronto, Canada

Jan 2025 to Apr 2025

Built recommendation system for 500+ community services → 40% better accuracy

03

University of Ottawa

Research Assistant · Ottawa, Canada

Oct 2024 to May 2025

Analyzed 100+ policy documents to study tech activism impact

Things I built

Research papers, ML experiments, and a growing list of side projects.

⭐ Featured

The Lab — Daily Games

A personal playground of daily browser games: Billboard Ranker (drag songs into chart order), Timeline (arrange historical events oldest-to-newest), and more. Built with Next.js, streaks, and share cards.

Next.jsTypeScriptTailwind CSSVercel
⭐ Featured

World Cup 2026 Predictor

Full-stack bracket prediction app for FIFA World Cup 2026. Pick every match from group stage through the final, share your bracket, and compete on a leaderboard. Built with Next.js 15, Supabase, and real-time social features (reactions and comments).

Next.jsSupabaseTypeScriptTailwind CSSVercel

other projects

Predictive Genomics ML Pipeline

Published ML pipeline (Elsevier, 2025) for Mendelian genetic-disorder recurrence risk prediction using clinical history and pedigree features. Tuned Gradient Boosting with SMOTE and bootstrap CV to prioritize sensitivity — 99% CV accuracy.

PythonScikit-learn+3

Prefiguration Lab

Website for a research lab studying prefigurative politics and technology activism. Built to be fast, accessible, and low-maintenance for a non-technical team.

Next.jsTailwind CSS+1

Fake News Detection

Passive Aggressive classifier on a tweet corpus to flag misinformation. Packaged preprocessing and inference for real-time classification at 90%+ accuracy.

PythonNLP+2

Stock Market Prediction

Time-series prediction experiments combining preprocessing and regression models to forecast stock movement.

PythonPandas+2

Let's build something

Open to collaboration, discussions, or just a conversation about AI.

Email

Send direct email

LinkedIn

linkedin.com/in/adithiyanrajan

GitHub

github.com/Adithiyan

Based in Ottawa, Canada. Available for collaboration and discussion. Usually reply within 24h.