Open to research collaboration and interesting projects

Kasidet Modem

Kasidet 'Modem' Saisudsawat

PhD Student, Computer Science

Kasidet Saisudsawat

Dept. of Computer Science and Information Technology, Faculty of Science, Naresuan University, Phitsanulok

I work on computer vision and deep learning for medical images, and I build the systems around research: sensors, RAG, and the infrastructure that keeps them running.

Email me

Research fields

  • Computer Vision
  • Deep Learning
  • Vision Transformers (ViT)
  • Convolutional Neural Networks (CNN)
  • Medical Image Analysis
  • Digital Pathology
  • IoT & Embedded Systems
  • Sensors & Instrumentation
  • Technology for Plant Science Labs
  • RAG & LLM Systems
  • DevOps & Infrastructure
Research projects
3
Company systems, 2025 to 2026
3
Publications
2
Building software since
2021

Research [3]

Research

From sensor systems for plant science labs to deep learning for medical image analysis. Published work links to the paper.

PhD, in progress

Deep learning for medical image analysis (computational pathology)

A generic pipeline of the techniques, not of the study or its results.

PhD research in computer vision and deep learning for medical image analysis. I train CNN and Vision Transformer models and adapt pathology foundation models for segmentation and detection in very large digital pathology images.

  • Computer Vision
  • Deep Learning
  • Vision Transformer (ViT)
  • Foundation Models
  • Semantic Segmentation
  • Object Detection

Details after publication.

We added IoT to a lab's plant tissue-culture temporary immersion system using Arduino, Raspberry Pi, Firebase and a webcam, so researchers can change feeding schedules and check each feed from a web page.

  • Internet of Things (IoT)
  • Temporary immersion system (TIS)
  • Micropropagation
  • Arduino
  • Raspberry Pi
  • Remote monitoring

Best Thesis Award (1st place), Naresuan University, 2021

ICIC Express Letters Part B, 2022 · DOI 10.24507/icicelb.13.11.1123

Read summary

Work [3 systems, 2025 to 2026]

Selected work

Full-stack and DevOps developer at a company whose name stays confidential. I built, deployed and ran these three systems; the rest are listed below them. Oct 2025 to Aug 2026

Company system · Full stack and DevOps

Multi-engine OCR platform with AI quality checking

Two OCR engines read the same Thai document, and an LLM checks and corrects the result.

  • Three modes: fast Thai OCR, layout boxes, or a 4-pass high-accuracy run
  • Two engines cross-check each other, and an LLM judges the unsure regions
  • Async API with Google login, API keys and usage quotas
  • DeepSeek-OCR
  • Typhoon-OCR
  • LLM judge
  • FastAPI
  • vLLM
  • Kubernetes
Open details
Clientor n8n workflowFastAPI hubauth, API keys, tasksDeepSeek-OCRpass 1: layout, textTyphoon-OCRpass 2: verify regionsvLLMmodel endpointsLLM judgepasses 3-4Resulttext + bounding boxes

A constraint solver builds a hospital ward's monthly nurse roster from staffing rules and preferences.

  • Builds the whole month's roster with OR-Tools CP-SAT
  • Keeps staffing and rest rules, then meets preferences where it can
  • Nurses request leave and swaps; the ward head approves
  • OR-Tools CP-SAT
  • FastAPI
  • React 19
  • PostgreSQL
  • Docker Compose

Contract systems that read scanned and digital contracts and answer questions with RAG and GraphRAG.

  • Reads digital and scanned contracts, with OCR for scanned pages
  • Splits the text by Thai legal structure and keeps page numbers
  • Answers questions with RAG, GraphRAG and a tool-calling agent
  • RAG (pgvector)
  • GraphRAG (Neo4j)
  • LLM agent + tool calling
  • FastAPI
  • Cohere rerank
  • Celery + Redis

Other and earlier work

  • 2025 to 2026
    Also built at the same companyA Thai legal Q&A assistant, a document RAG pipeline on n8n with hybrid search and citation checks, document and thesis archives with AI search, and Ansible deployment automation.
  • 2023 to 2024
    Full Stack Developer, O.H.M Digital Life Co., Ltd.Insurance platform, e-commerce, car inventory and a hotel CMS, in Next.js, Laravel and MySQL.
  • 2021
    Intern, Western Digital TechnologiesYield-analysis visualization and a secure data-transfer protocol.

Freelance [3]

Freelance projects

Web systems I built for faculties at Naresuan University as a freelance developer, from the first requirements to deployment.

2025 to 2026 · Solo developer and DevOps

SCIPUB: research publication platform for the Faculty of Science

Researcher profiles, publications and dashboards for a whole faculty, with an AI semantic search.

Client: Faculty of Science, Naresuan University

  • The data pipeline: Google Scholar harvesting with Scopus and SJR quartile lookups
  • University single sign-on (LDAP) and a role model for staff and administrators
  • The semantic researcher search on pgvector, with AI keywords
  • Next.js 15
  • PostgreSQL + pgvector
  • Prisma
  • LDAP SSO
  • OpenRouter (Gemini)
  • Python scraper
  • Docker

Smaller jobs

  • 2026
    POSN registration system, Faculty of Science, Naresuan UniversityQR-code payment through the Krungthai bank API; a k6 load test at 300 to 500 concurrent users with a capacity report.

Stack

Skills

AI and LLM systems

  • RAG
  • GraphRAG (Neo4j)
  • LLM agents and tool calling
  • Hybrid search (dense + BM25 + RRF)
  • Vector databases (pgvector, Qdrant)
  • OCR engines (DeepSeek-OCR, Typhoon-OCR)
  • Computer vision (PyTorch, CNN, ViT)
  • Reranking (Cohere)
  • Model serving (vLLM, LiteLLM, Ollama)
  • Constraint solving (OR-Tools CP-SAT)

Web and backend

  • FastAPI
  • React
  • Next.js
  • PostgreSQL
  • TypeScript
  • SQLAlchemy + Alembic
  • Tailwind CSS + shadcn/ui
  • Celery + Redis
  • Auth (JWT, Google OAuth, RBAC)
  • Laravel + MySQL

Infra and DevOps

  • Docker
  • Kubernetes
  • Ansible
  • n8n automation
  • Docker Compose
  • Nginx Proxy Manager
  • GitHub Actions CI/CD
  • Linux servers + SSH jump host
  • MinIO + PostgreSQL backups

IoT and sensors

  • Raspberry Pi
  • Arduino
  • Firebase
  • MPRLS pressure sensor
  • DHT11 temperature/humidity sensor
  • PSY1 leaf psychrometer
  • Webcam capture
  • Web apps for lab devices

Beyond the code

Awards, teaching and talks

Awards

  • Best Thesis Award (1st place)Naresuan University, 2021
  • 2nd place, Undergraduate Innovation ContestNaresuan University, 2022

Teaching

  • Teaching assistantSensing & Actuation for IoT, Computer Architecture

Talks

  • Conference presentationsICICIC 2023 in Japan, AJCC 2022

Publications

  • 2 papersICIC Express Letters Part B, from the bachelor's and master's research

Open to research collaboration and interesting projects

Have a project or a research idea?

For a research collaboration, or a system that needs building and running, email reaches me fastest.

kasidetdem@gmail.com