Marcus consistently demonstrates strong ownership and pays close attention to detail. He is highly responsive in applying customer feedback and consistently delivers on his commitments — his reliability and accountability build trust with stakeholders and contribute positively to overall team effectiveness.
Open to the right opportunity
Marcus Chong
Software Engineer — Artificial Intelligence
I build production LLM applications, retrieval pipelines, and multimodal AI systems — turning research into reliable, production-grade software.
7+ Years of
Experience
Experience
Scroll
About
AI that holds up
in production
I'm an AI engineer based in Fullerton, California, with 7+ years at Boeing building production LLM and AI systems in defense aerospace. My work spans LLM applications, retrieval, and multimodal sensor data processing — from low-level network systems early in my career to the production ML pipelines I ship today — with a focus on AI that actually works in the real world.
At Boeing, I've built LLM-powered tools grounded in enterprise data, benchmarked retrieval strategies for engineering documents, productionised a computer vision model with 98% detection accuracy that cut inspection reporting time 67%, and architected the team's 0-to-1 ML infrastructure. I have two U.S. patents pending in multimodal data processing and LLM inference optimisation, and my work won Best Paper at Boeing's largest internal AI conference. I care about the unglamorous details: robust pipelines, thorough testing, and documentation that teams can actually rely on.
Outside of work I run a home lab, contribute to open source, and stay sharp on the latest in systems design and applied ML. If it involves building something with real constraints and real stakes, I'm interested.
Tech I work with
LLM & Retrieval
- Prompt Engineering
- RAG
- Hybrid Retrieval + Reranking
- ColPali
- Docling
- Unstructured
- Gemma
- OpenAI/GPT
Vector & Embeddings
- Milvus
- ChromaDB
- Embedding Pipelines
- Semantic Search
Multimodal & CV
- YOLOv8
- OpenCV
- EasyOCR
- Tesseract
- Faster Whisper
- PyTorch
- scikit-learn
Deployment & MLOps
- vLLM
- CUDA
- On-Prem GPU
- Docker
- Kubernetes
- GitLab CI
- Model Versioning
Languages & Platform
- Python
- TypeScript
- JavaScript
- C#
- C++
- SQL
- AWS
- Linux
- REST/OpenAPI
Security
- ITAR/EAR Compliance
- Access Control
- Secure API Design
Recognition
Patents & awards
02
U.S. Patents Pending
Multimodal data processing and LLM inference optimisation — filed May 2026.
01
Best Paper Award
Boeing's largest internal AI conference (2025). Co-author of an internal research publication on AI-powered automation.
05
Conference Presentations
One delivered in 2025 and four delivered in 2026.
Experience
Where I've worked
01
Software Engineer — Artificial Intelligence
Boeing
Building production LLM applications, multimodal sensor pipelines, and the ML infrastructure behind them — applied research meeting production engineering in a safety-critical domain.
- Built production LLM-powered tools integrating models with internal systems, REST APIs, and enterprise data — evaluating frontier and self-hosted models (Gemma family, internal GPT-based service) per use case, with prompt engineering and grounding for reliable outputs in safety-critical settings
- Benchmarked retrieval strategies on engineering documents — Docling and Unstructured parsing, ColPali, and hybrid retrieval with reranking — and engineered embedding pipelines on Milvus (ChromaDB for prototyping)
- Architected a production multimodal pipeline over camera video, audio, imagery, and documents — frame extraction, YOLOv8 object detection, OCR, and Faster Whisper speech-to-text; patent pending on the processing techniques
- Built and productionised a computer vision model — 98% detection accuracy at 0.5–3s latency, cutting inspection reporting time 67% in live operations
- Established team-wide evaluation and testing standards — retrieval-quality test sets, output validation, pytest regression suites — and built the CI/CD, Kubernetes, and on-prem NVIDIA GPU infrastructure behind them: the team's 0-to-1 ML platform
- Built ITAR/EAR access-control middleware enforcing export controls across international boundaries — secure-by-design within DoD-relevant compliance constraints
- Mentored engineers and led cross-functional initiatives, delivering AI enablement across engineering and operations
PythonPyTorchYOLOv8RAGMilvusChromaDBvLLMCUDADockerKubernetesGitLab CIMongoDBAWS S3
02
Software Engineer — Network & Cyber Systems
Boeing
Focused on network systems engineering — designing messaging frameworks, simulation environments, and performance optimisation across complex network topologies.
- Deployed unified messaging framework combining UDP speed with TCP reliability
- Architected network simulation capabilities for testing software across diverse topologies
- Analysed network performance bottlenecks and optimised message delivery systems
TCP/IPUDP/IPActiveMQNetwork SimulationC++
Education
Where I studied
Master of Science in Computer Science
Georgia Institute of Technology
Bachelor of Science in Computer Science
Chapman University
Kind words