Open to opportunities

Timothy Lee

@timothylee3

Message

Senior Machine Learning Engineer building scalable LLM-driven retrieval, RAG, and multimodal search systems.

United States

Message

What I'm looking for

I’m looking to build real-time, user-aware search and personalization with agentic LLM workflows—owning retrieval quality, relevance, latency, and scalable cloud deployment in a team that ships.

I’m a Senior Machine Learning Engineer with 7+ years of experience specializing in search, relevance, ranking, and NLP-driven content understanding. I build scalable retrieval, RAG, and personalization systems using LLMs, embeddings, and multimodal models, with a strong focus on real-time, user-aware search experiences.

At Roblox, I built a LangGraph-based multi-agent system for real-time discovery, search, ranking, and personalized itineraries on scalable AWS microservices, and I developed a hybrid-retrieval RAG knowledge platform (BM25, TF-IDF, embeddings) with caching for low-latency answers. Previously at Apple, I engineered synthetic-data training for conversational models on SageMaker, built multimodal chat agents (including text-to-SQL, graph visualizations, and contract generation), and delivered measurable impact like a 45% improvement in sentiment analysis accuracy. Earlier at Workday, I designed an NLP resume-screening platform with spaCy NER and implemented document QA using vector databases, FAISS-based similarity search, and scalable asynchronous document pipelines.

Experience

Work history, roles, and key accomplishments

Current

Machine Learning Engineer

Current

Roblox

Jun 2025 - Present (1 year)

Built a LangGraph-based multi-agent system delivering real-time discovery, search, ranking, and personalized itineraries on AWS microservices. Developed hybrid RAG with fine-tuned embeddings for low-latency answers, built a LLaMA-3 legal search assistant, and implemented a SigLIP-powered VLM analysis pipeline with caching.

LangGraph Multi Agent Systems (MAS)RAG Microservices LLaMA Hybrid Retrieval (BM25 TF IDF Embeddings)VLMs (SigLIP)

Machine Learning Engineer

Apple

Nov 2020 - Jun 2025 (4 years 7 months)

Engineered synthetic data generation to train and fine-tune conversational models, deploying on SageMaker for real-time interactions with reduced latency. Enhanced multimodal real-estate and other user-facing chat experiences using LangChain/RAG and LLM/VLM workflows, including sentiment analysis pipelines that improved accuracy by 45%.

Synthetic Data Generation Sagemaker LangChain RAG Prompt Engineering Multimodal NLP (Whisper BERT)Sentiment Analysis Text To SQL Kafka DBT ETL Lang Chain

Machine Learning Engineer

Workday

May 2019 - Nov 2020 (1 year 6 months)

Designed a resume-screening platform using spaCy NER to streamline parsing of resumes and job descriptions. Built document QA and retrieval pipelines with LangChain, NLTK, and FAISS, including an asynchronous PDF processing workflow and MongoDB-backed storage for scalable similarity search and analytics.