Open to opportunities

Steven Chen

@stevenchen1

Message

Senior machine learning engineer building scalable LLM and RAG systems for real-world impact.

United States

Message

What I'm looking for

I’m looking to build production-grade LLM, RAG, and multi-agent systems with strong MLOps/LLMOps and responsible AI guardrails—on cloud-native stacks—where I can deliver measurable reliability, safety, and low-latency impact.

I’m a Senior Machine Learning Engineer and Data Scientist with 10+ years building scalable AI systems and deploying production-grade machine learning and generative AI solutions. I specialize in LLM systems, multi-agent architectures, retrieval-augmented generation, and cloud-native ML platforms, with a track record of delivering reliable, high-impact AI products across healthcare and fintech.

I lead end-to-end LLMOps and MLOps implementations—prompt versioning, evaluation workflows for factuality and safety, hallucination detection, and token/cost monitoring—while building robust safety and compliance guardrails. At Truveta, I developed fine-tuned clinical LLMs (SFT, PEFT/LoRA, RLHF), deployed hybrid RAG pipelines over large-scale EHR data, and engineered MCP-style tool orchestration for automated clinical research workflows.

Experience

Work history, roles, and key accomplishments

Current

Senior Machine Learning Engineer

Current

Truveta

Jan 2023 - Present (3 years 5 months)

Led clinical LLM development using SFT, PEFT/LoRA, and RLHF, improving factual accuracy and guideline alignment by 31%. Built LangGraph multi-agent workflows and hybrid RAG over large-scale EHR data, reducing research cycle time by 35% and improving answer relevance by 24%.

LLM Fine tuning PEFT LoRA LangGraph Retrieval Augmented Generation (RAG)Azure Machine Learning Triton Inference Server HIPAA PHI Safety Guardrails

Machine Learning Engineer

Hugging Face

Jul 2018 - Dec 2022 (4 years 5 months)

Built and fine-tuned transformer models (BERT/RoBERTa/DistilBERT) for NLP tasks, improving accuracy by up to 20% on benchmark datasets. Delivered low-latency inference with FastAPI and deployed embedding-based retrieval using Sentence Transformers and FAISS, improving retrieval precision by 25%.

Transformers PyTorch fastAPI Sentence Transformers Faiss Retrieval Augmented Generation (RAG)MLFlow Airflow

Machine Learning Engineer

Amazon

Jan 2015 - Jul 2018 (3 years 6 months)

Built demand forecasting models using DeepAR and LSTM time-series methods to improve inventory planning for volatile and long-tail retail categories. Implemented large-scale Spark/EMR pipelines for feature generation and backtesting, and added monitoring for data quality and forecast stability to reduce undetected model degradation.

DeepAR Anomaly Detection Apache Spark EMR AWS Rolling Window Training Backtesting Model Monitoring SQL Python

Software Engineering Intern

Amazon

May 2013 - Aug 2013 (3 months)

Built data preprocessing pipelines with Hive/EMR/SQL to aggregate product view and clickstream signals for Amazon search and homepage experimentation. Conducted exploratory data analysis and prototyped predictive models in Python/NumPy to estimate user re-engagement and purchase intent.

Hive SQL Python NumPy Click Stream Data predictive modeling Data Pipelines