Guide
Welcome to my personal website! Hi, I'm Guide, an ML Engineer and Full-Stack Developer. My passion lies in creating systems that address real-world challenges through ML/DL — from developing AI tools like chatbots and RAG systems to deploying them in production.
These days I'm also CTO at MR-Rehab and founder of SolveServe Group, where my team and I turn those systems into products people use.
With experience in web development, cloud platforms, and CI/CD pipelines, I can find the right solutions for your problems and turn them into functional products and services.
Chulalongkorn University
Computer Engineering & Digital Technology
Thai (Native)
English (Fluent)
Japanese (Conversational)
Technical Stack
- >Python, TypeScript, Rust
- >PyTorch, Scikit-learn
- >React, Next.js, FastAPI
- >Kubernetes, Docker, ArgoCD
- >Terraform, Cloudflare
Focus Areas
- >AI/ML Systems
- >Healthcare Technology
- >Cloud Infrastructure
- >Startup Operations
Languages
- >Thai (Native)
- >English (Fluent)
- >Japanese (Conversational)
Education
- >Chulalongkorn University
- >Computer Engineering & Digital Technology
- [1]On creating an English-Thai code-switched machine translation in medical domain
Findings of the Association for Computational Linguistics: EMNLP 2024 · 2024
Code-switched medical MT that keeps critical English terminology in Thai text — competitive with Google NMT and GPT-3.5/4, and strongly preferred by medical professionals in human evaluation.
co-author cited by 7 - [2]Decomposing Food Images for Better Nutrition Analysis: a nutritionist-inspired two-step multimodal LLM approach
IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · 2025
A two-step prompting framework for off-the-shelf MLLMs: deconstruct the dish into ingredients and portions, then compute calories and macros — no fine-tuning or food databases required.
co-author cited by 5 - [3]SICAR at RRG2024: GPU Poor's Guide to Radiology Report Generation
Proceedings of the 23rd Workshop on Biomedical Natural Language Processing (BioNLP) · 2024
Radiology report generation with a lightweight MLLM plus a 'First, Do No Harm' SafetyNet — LLM post-processing cross-verified by an X-ray classifier. 4th place on F1-RadGraph F1 in the RRG24 shared task.
co-author cited by 2