Skip to content
View paipeline's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report paipeline

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 supported. This note will be visible to only you.
Report abuse

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

Report abuse
paipeline/README.md

Alex (Pai) Peng

M.S. Electrical and Computer Engineering - AI/ML Systems @ Carnegie Mellon University


LinkedIn


About

Alex (Pai) β€” I and my team research, prototype, and ship ML systems.

  • Nodetree.io: Tree-structure LLM for parallel queries.
  • Prisma: Agentic Planning Abstraction Layer, building Workflow-as-Code for repeatable workflow generation.
  • Deepthink in MCTs: Developing a new way to udpate RL rewards based on Deepthink with Confidence and Monte Carlo Tree Search unsupervised iterations.
  • Language to Langgraph: automatic LangGraph construction via LLM as finite state machines.
  • T-Few: transfer learning with few-shot setups for low-label regimes.
  • Raman Spectrum Analysis: CNN + ordinal classification to detect peaks; separated PAHs vs. pesticides with high accuracy.
  • Low-Light Image Enhancement: benchmarked DL models for lab imaging. Submitted to CVPR 2026 ``

AWS

  • Solutions Architect – Associate
  • 🎯 Preparing Machine Learning – Specialty (cloud-native, scalable AI pipelines)

Work Experience

  • American Battery Solutions (Software Engineer Intern, 2025) β€” delivered code-intelligence and auto-doc portal (libCST + FastAPI) that cut manual documentation.
  • ISVision Technology (Software Engineer Intern, 2024) β€” implemented an AI-driven code reviewer over GitLab commits (LLaMA + RAG) to automate checks.
  • Computer Vision Center, Barcelona (Research Assistant, 2022–2023) β€” co-authored MILL low-light dataset and benchmark; ran SOTA baselines and prepared splits.

Projects (RIP)

  • MyMindy β€” Data Mining Your Notes: AI/NLP to extract & surface actionable insights from personal notes. β†’ https://www.Mymindy.net
  • Step.ai β€” AI↔AI networking (coffee chats for agents; autonomous resume exchange & interaction)
  • Vitaelia β€” AI Vision Necklaces (burned an Arduino πŸ”₯)

Languages I speak

  • Spanish: (Bilingual)
  • Chinese: (Bilingual)
  • English: (Advanced)

Popular repositories Loading

  1. deep-think-with-confidence deep-think-with-confidence Public

    This repo serves as an early adaptation that offers a threaded Openai DeepThink-with-confidence demo for research purpose before the official code release.

    Python 9

  2. JobPush.ai JobPush.ai Public

    Python 2 1

  3. NovelGPT NovelGPT Public

    NovelGPT: Creating high quality novels based on multi-agent system powered by vectorbase self-feeding data.

    Jupyter Notebook 1

  4. novel-GPT-v2 novel-GPT-v2 Public

    Jupyter Notebook 1

  5. ai2aidemo ai2aidemo Public

    Python 1

  6. concent-demo concent-demo Public

    A proof of concept for a mobile concentration tool that starts recording study/work time when the mobile phone is flipped over.

    TypeScript 1