Phattharanat Khunakornophat
Data Scientist/Developer
I have three years of experience as a Data Scientist and Developer. Most of my tasks are about providing analyses to support a strategy team. I gained experience in the field of spatiotemporal data and time series analysis whilst working on my senior project.
Experience
- Provide data analysis and visualisation for the strategy team
- Implement data scraper for social media platforms including Pantip (Thailand discussion forum), Twitter, YouTube and TikTok
- Analyse point of sale data including RFM analysis, cohort analysis and association rule learning
- Train a Thai Speech Emotion Recognition model using Transformer architecture, achieving 60% validation accuracy across five classes data
- Analyse and visualise survey responses using word cloud, n-gram and LDA topic modelling
- Research, design and develop backend systems and data pipelines for a proof-of-concept project centred in real-time people detection, face recognition using Django, FastAPI, DBSCAN, Airflow, Chroma
- Research and implement Retrieval-Augmented Generation and Model Context Protocol for Postgres and Jira
- Implement basic Bitbucket Pipelines to automate the deployment to an internal testing server, eliminating manual tasks
- Maintain and refactor Django starter template, adding TOTP and LINE login features, websocket-based notifications, migrating package manager from pip to uv
- Explore, containerise, scaffold project structure for backend frameworks including Go Fiber, Elixir Phoenix
- Develop a proof-of-concept vehicle detection system using TimescaleDB, Kafka, YOLO, and RF-DETR
- Configure and manage cloud instances, including setting up environments, installing dependencies, and managing Nginx and Cloudflare for web infrastructure
- Research and set up an internal shadcn component registry to enable reusable UI components across projects
Projects
A collection of snippets of code for solving problems bundled as python package that can be installed from github
Implement the Wordle puzzle in Haskell language to explore functional programming style
Solve the Wordle puzzle optimally by using statistics/information theory
- Developed a backend server for climate analysis tasks: trend analysis (Mann-Kendall test), spatiotemporal data visualisation
- Performed and evaluated performance of climate bias correction methods
- Implemented basic machine learning models: linear regression, linear classification, K-mean, Naïve Bayes, bias-variance trade-off
- Explored and visualised NILM (Non-Intrusive Load Monitoring) dataset
Activity
- Provided a retrospective analysis of a chart and proposed improvements
- Highlighted some great data visualisations in history
Skills
Education
Interests
- Data visualisation
- Table Tennis
- Listening to Sawano Hiroyuki's music
- Functional Programming, Haskell, Elixir
- Personal Knowledge Management, Zettelkasten, Evergreen Notes
Now
Welcome to my resume! I want to use this section to tell you about what I
am doing now.
(Last Updated: 2024-04-14)
I am currently interested in Functional Programming and Monad. I enjoy drinking Earlgrey tea and specialty cofee.
I am also suffering from Gastritis, Conjunctivitis, Dry Eye, De Quervain's tenosynovitis.
I have tried many note-taking apps such as Roam Research, Craft, Logseq, Obsidian, Tana, Heptabase. I am practising taking notes with the Zettelkasten/Evergreen notes method but I am still struggling with it.