“Every good gift and every perfect gift is from above, coming down from the Father of lights.” — James 1:17
Hi, I’m Jae 👋
Machine Learning Engineer & AI Specialist with 8+ years of experience delivering production-grade ML systems. I specialize in LLM fine-tuning, RAG architectures, and MLOps at scale.
Currently a Research / Machine Learning Engineer at Meta. Across my career, I’ve led initiatives that have:
| Impact | Metric |
|---|---|
| 💰 Cost Savings | $250K+ annually across 15+ projects |
| ⚡ Efficiency | 38% reduction in development time |
| 📈 Revenue | $1.5M+ additional annual revenue |
| 🎯 Accuracy | 15% improvement in predictive models |
What I Do
🤖 LLM & RAG Systems
Building production multimodal RAG pipelines with Llama, GPT, and Gemini. Designing data-centric chatbots that scale.
⚙️ MLOps & Platform Engineering
Architecting Kubernetes-based ML platforms with automated CI/CD, model monitoring, and continuous learning feedback loops.
🎙️ Voice AI & Agents
Engineering low-latency voice agents with streaming ASR-to-LLM pipelines, achieving sub-4s response times at $0.05/conversation.
Experience
Research / Machine Learning Engineer
Meta · Jan 2026 – Present
- More details coming soon.
Machine Learning Engineer | AI Engineer
Walmart Global Tech · Oct 2023 – Jan 2026
- Led design and rollout of standardized MLOps template integrating multimodal RAG and data-centric chatbots (Llama, GPT, Gemini-Pro)
- Established continuous learning feedback layer to monitor metrics, capture inference data, and automate retraining
- Impact: Cut development time by 38%, improved team efficiency by 30%, achieved $250K+ annual savings across 15+ production projects
Data Scientist
Walmart Global Tech · May 2021 – Oct 2023
- Architected Kubernetes-based automated pipeline integrating model logic, monitoring, and CI/CD with testing and code reviews
- Transformed model adoption from 8% to 14%, improved predictive accuracy by 15%, driving $1.5M additional annual revenue
- Eliminated critical downtimes, reduced manual effort by 40%, cut cloud costs by 20% — saving $120K annually
Machine Learning Engineer Lead
Mozzine · Sept 2020 – Feb 2021
- Built large-scale, high-quality dataset via custom scraping and cleaning pipelines with novel reconstruction algorithms
- Improved core model accuracy by 20%, driving 5% boost in customer conversions and enabling higher-value pricing tier
Featured Projects
🎙️ Voice-to-Voice Health Insurance Agent
Real-time voice AI for health insurance inquiries
- Engineered streaming ASR-to-LLM pipeline with multimodel routing (GPT, Llama3, Qwen, Gemma) and LoRA-based domain adaptation
- Results: p95 latency of 2-4s, 92% QA accuracy, 80-88% multi-turn task completion
- Cost: Only $0.05-$0.35 per five-turn conversation
🧠 KB Agent (Knowledge Base Assistant)
Enterprise multimodal knowledge agent
- Designed modular RAG architecture with dual embeddings (text + CLIP), HNSW + IVF vector search, and validity-gated answer generation
- Results: 60% reduction in lookup time, answer accuracy improved from 32% to 92%
- Scale: Supported 10K daily users, driving $53K in revenue
Technical Skills
Languages & Frameworks
Python · C++ · Java · Flask · FastAPI
AI / ML / Deep Learning
Claude · GPT · Llama · Gemini · Qwen · LangChain · LlamaIndex
TensorFlow · PyTorch · Keras · Scikit-Learn · OpenCV
Hugging Face PEFT · LoRA · Whisper
MLOps & Cloud Platforms
Kubeflow · Vertex AI · MLflow · Databricks · Airflow
Azure · GCP · AWS · Kubernetes · Docker · Git
Data & Storage
PostgreSQL · MySQL · BigQuery · Redis · Cosmos DB
Spark · Dask · Ray · Milvus · Chroma
Education
🎓 M.S. Robotics & Autonomous Systems (AI & CS concentration)
Arizona State University · GPA 4.0/4.0 · Dec 2020
Awards & Recognition
| Award | Year |
|---|---|
| 🥇 1st Place — AWS | State Farm Hackathon | 2020 |
| 🥇 1st Place — MESCON | 2014 |
| 🎖️ Engineering Graduate Fellowship — ASU | 2019-2020 |
| 📚 Academic Scholarship — UE | 2010-2015 |
Publications
📄 Scan & Go — US Patent No. 20240029119 (USPTO)
📄 Production And Measurement Of Four Degree Of Freedom Photonic States — Semantic Scholar
Let’s Connect
I’m always open to discussing ML engineering roles, AI architecture consulting, or collaboration opportunities.
📬 Get in Touch
Support This Blog ☕
This blog is my open notebook — where I write to remember, to share freely, and hopefully to save someone else the hard way. As a Christian, I believe every good gift comes from above (James 1:17), so I try to give back what I’ve been given: honest, ad-free content. If something here has helped you, a coffee keeps me writing.
💝 Ways to Support
Every contribution, no matter how small, is deeply appreciated — and goes right back into more notes, tutorials, and deep-dives for the community.
About This Blog
“It is easily forgettable that we are running into death from birth.”
This blog is my space for reflections, technical deep-dives, and lessons learned from 8+ years in the trenches of production ML. I share what I wish I had known earlier — the failures, the wins, and everything in between.
Topics I write about:
- 🧠 LLM fine-tuning & prompt engineering
- 🔧 MLOps best practices & war stories
- 📊 System design for ML at scale
- 💡 Career advice for ML engineers
Happy to be criticized to mature. Iron sharpens iron.