LLM, SLM, Multi-agent systems, llama.cpp, Gemma, LFM, RAG, AI Eval, DSPy, LlamaIndex, AWS Bedrock, Google Vertex AI, MLOps, Whisper, YOLO, Ray, TensorFlow, Pandas, NumPy, PySpark
- +6596325822
- Singapore
- info@maxongzb.com
Summary
Solution Architect with nearly 10 years across AI/ML, full-stack engineering and startup leadership. I design LLM, RAG and multi-agent systems, with a focus on data foundations, evaluation and production-readiness decisions. At ASUS I owned evaluation metrics for an ASR + LLM translation tool (80% domain-quality target met) and advised on architecture across 4 AI innovation projects. At Airbus I did the data lineage analysis and wrangling behind RAG proofs-of-concept on Palantir Foundry. As CTO of Codesurance I led delivery across multiple client engagements, including an event system serving 10,000+ participants. Pursuing Georgia Tech OMSCS (AI). PyCon Malaysia 2023 keynote speaker.
Work Experience
Asus
2024-122026-03
Software Design Team Lead
- Owned the evaluation framework for an open-source ASR + LLM Chinese/English dictation and translation tool, tracking CER, WER, latency and CPU/GPU/RAM usage. Met the 80% domain-specific translation quality target before handover to a business unit for scaling.
- Assessed AI/ML frameworks and technologies across 4 innovation projects, producing decision matrices and technical comparison documentation that informed architecture selection and investment planning.
- Provided technical advisory on system architecture, API contracts and integration patterns for a consumer innovation project, translating requirements from a cross-functional team (UX, industrial design, engineering) into a technical implementation roadmap.
- Conducted technical feasibility assessments with project managers and business stakeholders across Singapore and Taiwan, surfacing implementation and scope risks before development began.
- Contributed frontend architecture (React, TailwindCSS, Vite) to 3 AI innovation projects.
Airbus
2023-122024-06
Digital Solution Engineer
- Performed data lineage analysis and data wrangling across Skywise datasets, establishing the data foundation for RAG POCs and digital twin work.
- Evaluated open-source LLM frameworks (LlamaIndex, DSPy) for the RAG POC.
- Worked with APAC stakeholders to turn business problems into scoped POC requirements.
Upskillcycle Pte Ltd
20212024
Founder
- Founded an inclusive tech education platform targeting career switchers and persons with disabilities
- Secured Singapore government funding via Enterprise Singapore's SG Startup Founder Grant.
- Architected and built the full platform using Django, Bootstrap, Websockets, Celery, Redis, and Django Channels, scaling to 5,000+ users through Mentoring.SG.
- Placed Top 28 semi-finalist at Entrepreneur World Cup Singapore 2021 and 3rd place at Alibaba Cloud SUSS Demo Day 2021.
- Drove end-to-end product, technical, and community strategy as sole technical founder from MVP to grant-funded growth
Codesurance Pte Ltd
20192023
Chief Technology Officer (CTO)
-
Defined engineering practices and code review standards across delivery teams.
- Architected and delivered full-stack solutions across multiple client engagements spanning event management, marketing automation, HR tech, and mobile applications for startups and enterprises
- Built an Account-Based Marketing automation pipeline using Selenium, Scrapy, AWS EC2, Pandas, and Flask, reducing manual prospecting effort for clients
- Delivered a check-in and alcohol redemption system handling 10,000+ event participants, integrated with Twilio, Mailgun, and AWS for real-time operations
Microsec
20172019
Full Stack Developer
- Developed the IoT management dashboard to visualise data or manage IoT devices using AngularJS, Bootstrap, CSS3, Flexbox, Django, Flask.
- Building public, internal & external RESTful APIs for clients, developers or IoT devices with Django, Flask, Docker, Postman, Pandas, Numpy, Nginx, Gitlab
- Front-end & Backend integration of our products or services
- Kickstart and build MicroSec's Design System with Figma
Agent Swarm at the Edge (Personal) | github.com/steelwolf180/edge-agent-swarm
Python, DBOS, PostgreSQL, llama.cpp, Gemma, LFM, OpenTelemetry, Docker
- Designed and built a five-agent pipeline (Researcher, Architect, Scribe, Critic, Judge) that turns architecture specs into C4 diagrams, ADRs and scored critiques. It runs fully on-device on CPU-only hardware with small models, so specs never leave the machine.
- Orchestrated with DBOS and PostgreSQL for durable workflow state, versioned artifacts and a human approval gate before anything is persisted.
- Added OpenTelemetry tracing (Arize Phoenix) and pytest integration tests running real end-to-end inference on the approve and reject paths.
- Benchmarked against a 5-minute performance target, found it unachievable on CPU-only hardware, and revised it to under 15 minutes (met in testing).
Education
Georgia Institute of Technology
20242029
Master of Science, Computer Science
Specialisation in AI