Generative AI Specialist based in Toronto, Canada — 8+ years building in AI, NLP, and Deep Learning.

I work on large language models (LLMs), natural language processing, and applied machine learning — from research and fine-tuning transformers to shipping production AI systems. Currently building at Behavox.

• 🔭 Focused on: LLMs, retrieval-augmented generation (RAG), transformer fine-tuning, and AI automation
• 🌱 Interests: NLP, deep learning, document intelligence, and agentic AI tooling
• 📍 Toronto, Canada
• 💼 LinkedIn: in/rakshit-makan
• 🌐 Website: rakmakan.github.io
• ✍️ Writing: Medium @rakshitmakan

Python · PyTorch · Hugging Face Transformers · LLMs · RAG · NLP · Deep Learning · ONNX · TypeScript

• HITL Document Classification — Reference implementation of my master's thesis (Dalhousie University, 2022): a human-in-the-loop system for classifying multi-page scanned administrative documents, using a two-step Sentence-BERT classifier with incremental retraining. 🎓
• Clustering-with-BERT — Document clustering models built on BERT to efficiently process and organize large corpora of publications. ⭐ 17
• LLM_Projects — A modular Python application for machine learning model training and inference, supporting both Hugging Face and ONNX model formats.
• claude-job-hunter — A Claude Code plugin for end-to-end job application automation: brainstorm, discover, filter, tailor resumes, and auto-apply.

• I Turned My Job Search Into a Pipeline (and Let an Agent Run It) — how I built the claude-job-hunter plugin, and what applying to 60 roles in one night taught me about agents.
• Anatomy of a Claude Code Plugin: How Claude Job Hunter Works Under the Hood — the architecture behind it: skills, hooks, MCP, and a files-as-state pipeline.

I care about turning research-grade NLP and generative AI into things people actually use — clean pipelines, well-scoped models, and tooling that removes friction.

Rakshit Makan — Generative AI / NLP / Deep Learning Engineer.