About

About

About Me

AI/ML Engineer with 8+ years of experience building production AI systems and data infrastructure. I specialize in taking AI from prototype to production — from fine-tuned model serving to agent systems that operate on real business data.

Professional Experience

Principal Machine Learning Engineer — Magnify (2024-2025)

  • Built semantic data layer and ML evaluation pipeline for customer success automation platform
  • Designed agent reliability framework with tool-limit middleware and LangSmith eval pipeline
  • Led canonical data model initiative unifying fragmented data sources

Tech Lead, ML/AI Platform — AstrumU (2019-2024)

  • Led 6-person AI/ML team building talent intelligence platform (entity resolution at scale)
  • Architected multi-model serving with Ray Serve + vLLM: 10K+ entities/day online, 1M+ daily batch
  • Fine-tuned Qwen-2.5 with LoRA SFT for structured extraction, achieving 1-2 orders of magnitude cost reduction vs API
  • Built event-driven data pipelines with Kafka CDC + Dagster, replacing Databricks with unified PostgreSQL platform

Independent Consulting — Haishi Optoelectronics (2025-present)

  • Solo 0→1 AI agent for manufacturing ERP: schema design, RAG, SQL tools, ERP write-back, frontend, deployment
  • 6,600+ customers synced, automated reports, lead discovery pipeline
  • Production system with real users — not a demo

Education

Ph.D., Industrial & Systems Engineering — University of Washington

Publications

  • Gong, J., & Liu, S. (2023). Partially observable collaborative model for optimizing personalized treatment selection. European Journal of Operational Research, 309(3), 1409-1419. doi:10.1016/j.ejor.2023.03.014
  • Gong, J., Simon, G.E., & Liu, S. (2019). Machine learning discovery of longitudinal patterns of depression and suicidal ideation. PLOS ONE, 14(9): e0222665. doi:10.1371/journal.pone.0222665

Patents

  • Wray, A. J., Pedersen, K. O. P., Cai, X., & Gong, J. (2024). Predictive learner score (U.S. Patent No. 11,922,332).
  • Wray, A. J., Pedersen, K. O. P., Cai, X., & Gong, J. (2024). Predictive learner recommendation platform (U.S. Patent No. 11,928,607).

Technical Skills

  • AI/ML: LLM, AI Agent, RAG, Claude API, OpenAI API, vLLM, Ray Serve, PyTorch, Scikit-learn
  • Data: dbt, Dagster, Kafka, Spark, PostgreSQL, pgvector, Snowflake, DuckDB
  • Infrastructure: Docker, Kubernetes, FastAPI, AWS, GCP
  • Frontend: React, TypeScript, Next.js, Tailwind CSS
  • Languages: Python, JavaScript/TypeScript, SQL, R

Contact