Reliable LLM Systems

Hi, I'm Pete Pittawat Taveekitworachai

Researcher and builder shaping practical reasoning workflows for teams adopting large language models.

From Null-Shot Prompting (EMNLP 2024) and Prior Prompt Engineering (EMNLP 2025) to the Typhoon T1 multilingual reasoning suite, I translate frontier research into tools that ship.

Pete's profile picture

Current focus

Advancing reinforcement learning pipelines that strengthen reasoning fidelity.

Ship-ready builds

See the evaluation stacks and agent tooling delivered to teams.

Explore projects

Research notes

Read experiments, failure digs, and applied prompting patterns.

Dive into the blog

Talks & workshops

Watch practical walkthroughs from conferences and private sessions.

Listen to talks
Impact in Practice

About Me

I’m Pete (Pittawat Taveekitworachai) — a researcher focused on large language models and prompt engineering. My work explores how to make LLMs reliable, practical, and easy to use in the real world.

I bridge unconventional ideas to fundamental and applied research. I share what I learn through my blog, publications, and talks.

Reliability & RLHF

Reinforcement learning for advanced reasoning—Prior Prompt Engineering (EMNLP 2025) and Typhoon T1 anchor trustworthy multilingual agents.

Prompt Engineering

Null-Shot Prompting (EMNLP 2024), FinCoT, and structured playbooks make chain-of- thought dependable for high-stakes workflows.

Evaluation

BenchING (IEEE ToG 2025) and ChatGPT4PCG benchmark structured outputs for agents and games with reproducible scores.

Applications

LLM deployments across games, medical triage reasoning, and smart-car ADAS copilots where latency and safety matter.

Learn more about me

Research Focus

Where current publications concentrate and the systems I’m shipping next.

  • Reasoning models & RL

    Prior Prompt Engineering (EMNLP 2025) and Typhoon T1 advance RLVR/RFT pipelines while improving Thai reasoning fidelity.

  • Prompt engineering

    Null-Shot Prompting (EMNLP 2024) and FinCoT blueprint structured chain-of- thought for domain experts and analysts.

  • Evaluation & benchmarking

    BenchING (IEEE ToG 2025) and ChatGPT4PCG track structured outputs, levels, and agents across releases.

  • Applications

    Game storytelling, medical triage reasoning models, and smart-car ADAS copilots translate research into products.

Professional Highlights

Translating research into community resources, talks, and tooling.

  • Publications & writing

    First-author publications at IEEE ToG 2025 and EMNLP 2025 main track on dependable LLM systems.

  • Talks & workshops

    Sharing EMNLP and CoG findings through invited talks, workshops, and community labs.

  • Open-source tooling

    Typhoon T1 releases and BenchING tooling help teams ship reliable, open LLM workflows faster.

Proof of Work

What I publish and share with the community

Regular writing, peer-reviewed research, and talks that turn LLM research into approachable practice.

Latest Writing

Fresh experiments and field notes

Short reads on evaluation, prompting techniques, and the practical side of running LLM systems in production.

Browse all articles