Skip to content
View larrygrpolanco's full-sized avatar
3 followers · 4 following

Achievements

Achievement: Pull SharkAchievement: YOLO

Achievements

Achievement: Pull SharkAchievement: YOLO

Block or report larrygrpolanco

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
larrygrpolanco/README.md

Hi, I'm Larry

I build tools for education and research, mostly at the intersection of language, data, and software. By day I work on the data side of cancer research at Moffitt Cancer Center. On the side, I run Chaone Labs where I build apps for language researchers and educators.

What I'm working on

  • ITA Trainer: A voice-based practice tool for international teaching assistants. You talk to an AI student, then get coaching feedback on your interactional skills. Built in collaboration with education researchers. (Next.js, LiveKit, OpenAI Realtime API)

  • 한국어 학습 플랫폼: A web platform that supplements the KLEAR Integrated Korean textbook with LLM-powered interactive practice. (SvelteKit)

  • LLM Temperature Consistency: Replication package for a study on how temperature settings affect LLM reliability in rhetorical move-step annotation. Designed so non-programmers can run it. (Python, OpenAI API)

  • Pathology LLM Data Extraction: Evaluating open-source LLMs for zero-shot thyroid cancer pathology abstraction at Moffitt Cancer Center. (Python, LLM APIs)

Past projects

  • LIN2002 Grading Assistant: An AI grading tool for a linguistics and film studies course. It learns from real instructor feedback using few-shot examples and outputs scores and written feedback in the instructor's voice. (SvelteKit, Google Gemini)

  • The Office Agent: A virtual office simulation with AI agents that perceive, plan, reflect, and interact. Based on the "Generative Agents" paper by Park et al., ported to a simpler architecture. (Python, Phaser, FastAPI)

  • Fulbright Planner: An LLM-powered lesson planning tool I built while coordinating a Fulbright program in Taiwan. Tested with ~50 colleagues and refined based on actual usage logs. (Python, Svelte, Claude API)

  • SABR Automated Coding: A prototype that uses LLMs to do qualitative research coding on educational transcripts. Validated against expert human coders. (Python, LLM APIs)

  • YASEC: My first real coding project. A 58-million-word corpus built from YouTube academic transcripts to study the vocabulary demands of online lectures. (Python)

Pinned Loading

  1. standard-chinese-platform standard-chinese-platform Public

    A digital revival of the FSI's "Standard Chinese" course with a nostalgic cassette aesthetic.

    Svelte

  2. ETA-lesson-planner ETA-lesson-planner Public

    An LLM-powered lesson and workshop planning tool designed specifically for Fulbrighters in Taiwan

    Svelte 1

  3. ita-trainer ita-trainer Public

    TypeScript

  4. llm-temperature-consistency llm-temperature-consistency Public

    Python

  5. AutoAnki-SLA AutoAnki-SLA Public

    Python

  6. TCGA-THCA-pathology-llm-abstraction TCGA-THCA-pathology-llm-abstraction Public

    Zero-shot LLM extraction of thyroid cancer pathology variables; prompt, data, and results from a 7-model evaluation on TCGA-THCA reports.

    Python