Leon Martin is an AI Full-Stack Developer with over 6 years of experience, specializing in integrating large language models into real applications using Hugging Face Transformers and LangChain. He builds end-to-end intelligent systems including document analysis tools, RAG (retrieval-augmented generation) platforms, and chat-based automation. Leon also manages full-stack responsibilities, including backend APIs and frontend integration for LLM apps.
He is especially skilled in token optimization, model fine-tuning, and deployment on modern cloud infrastructures. Outside work, Leon contributes to open-source LLM tooling and experiments with new Hugging Face models and embeddings.
Skills
- Hugging Face
- LangChain
- OpenAI
- Python
- Docker
- FastAPI
- Redis
- Weaviate
Experience
- 2023 - Present
AI Applications Engineer à Humanloop (humanloop.com)
Leon worked on customer-facing LLM tools at Humanloop, building prompt-based workflows and supporting LangChain agents for use in customer support and document summarization tools. He contributed to prompt templating systems, fine-tuned models via Hugging Face, and deployed inference pipelines using vector databases for context-aware completions.
- LangChain
- Hugging Face Transformers
- Python
- OpenAI API
- PostgreSQL
- Redis
- 2021 - 2023
Machine Learning Developer à Cohere (cohere.ai)
Leon built AI-driven content generation tools and semantic search engines using transformer models. He handled model configuration, evaluation metrics, and built RAG pipelines using Hugging Face embeddings with LangChain integration. He also maintained backend infrastructure for model inference and prompt caching.
- Python
- Hugging Face
- LangChain
- FastAPI
- Weaviate
- Docker
- 2019 - 2021
Data Engineer à Zeta Alpha (zetalpha.com)
At Zeta Alpha, Leon worked on backend data pipelines for research indexing and content aggregation. He was responsible for processing large sets of academic metadata, developing keyword tagging mechanisms using traditional ML models, and building internal tools for search relevance testing. He also contributed to the frontend filtering logic for document categories and author clusters.
- Python
- ElasticSearch
- Scikit-learn
- VueJS
- PostgreSQL
- 2017 - 2019
Software Engineer à LightTag (lighttag.io)
Leon worked on the development of annotation tools for text data. His responsibilities included building frontend features for labeling workflows, creating backend logic for dataset management, and adding basic rule-based entity detection to help annotators tag named entities more efficiently.
- Python
- Flask
- MongoDB
- Jinja2
- JavaScript
Education
- 2017
Bachelor of Science in Computer Science à Ho Chi Minh City University of Science (VNUHCM - University of Science)
