Hey, I'm
|
AI Engineer with expertise in LLMs, Agentic AI, RAG Systems, and MLOps. I build intelligent systems that bridge research to real-world application.
About Me
Hello! My name is Surakiat Kansa-ard — a passionate AI Engineer from Thailand.
I specialize in Large Language Models (LLMs), Agentic AI, and RAG Systems, ensuring robust AI solutions and scalable ML pipelines. I combine deep technical expertise with a problem-solving mindset, focusing on bridging the gap between research and real-world deployment.
My passion lies at the intersection of Artificial Intelligence and Engineering Excellence. I thrive on building intelligent systems that solve complex problems, from multi-modal RAG applications to end-to-end MLOps pipelines.
I value Continuous Learning, Collaboration, Innovation, and Growth Mindset above all else. These principles guide both my personal development and how I approach building AI systems that deliver meaningful impact.
Technical Skills
Technologies I use to build AI solutions
Generative AI & LLMs
Hugging Face
Ollama
OpenAI
DeepSeek
Llama
Qwen
Gemma
Gemini
Claude
DeepEval
Langfuse
FastMCPProgramming
Python
SQL
C/C++ML/DL Frameworks
PyTorch
TensorFlow
scikit-learn
pandas
numpyVector Databases & Embeddings
FAISS
Qdrant
Chroma
RedisVL
PostgreSQLMLOps
MLflow
DVC
Model Versioning
Experiment TrackingComputer Vision
OpenCVCLIP
EasyOCR
Image ClassificationData Engineering
ETL Pipelines
Airflow
Kafka
Data Cleaning
PyThaiNLPBackend & Deployment
FastAPI
Docker
REST API
GitHub ActionsVisualization
Streamlit
Gradio
Matplotlib
Plotly
GrafanaCloud & DevOps
Git
GitHub
DockerFeatured Projects
High-quality AI systems with verified metrics
Agentic Research System
A powerful, multi-agent AI research system built with LangGraph, RedisVL, and PostgreSQL. Features intelligent routing, semantic caching, FastMCP integrations, and Human-in-the-Loop (HITL) checkpoints.
Multi-modal RAG Shop Assistant
AI-powered product search system with multi-modal inputs (text/voice/image). Features zero hallucination with strict anti-hallucination protocols and Thai language support.
Multi-Stage Retrieval RAG System
Bilingual (Thai/English) RAG system for corporate policies featuring hybrid search and cross-encoder reranking.
Uber Delivery Time Prediction MLOps
End-to-End MLOps pipeline for real-time delivery prediction with automated retraining and streaming inference.
OCR Document Classification System
AI-powered invoice & receipt analysis system with automated document processing using Computer Vision and NLP for text extraction and classification.
Animals Detection with Deep Learning
Real-time animal detection system using YOLOv8n for 7 COCO classes with live counting, InfluxDB storage, and interactive dashboard.
Cyber-RAG: Cybersecurity Assistant
Specialized RAG system for cybersecurity documentation with zero hallucination from OWASP and MITRE ATT&CK.
Work Experience
IoT Solution Engineer Intern
Swift Dynamics
Developed Smart Camera Security solutions for industrial applications. Implemented real-time anomaly detection using OpenCV, reducing human errors by 75%.
Certifications
Professional credentials and achievements