AI engineer & researcher

I build AI systems
and the tools that
make them useful.

I’m Kipngeno Koech, an AI engineer and researcher. I work across machine learning, AI agents and software, turning ideas into tools people can use.

GitHubLinkedInKigali, Rwanda
Carnegie Mellon AI graduate · Kigali, Rwanda
Portrait of Kipngeno Koech
Research interestsVLAs · Agents · WAMs
Machine learningAI agents & developer toolsResearch & engineering

Selected work

Things I’ve built.

All projects
Developer tools

Weightless

Understand a model before downloading its weights. Explore architectures and memory requirements in the browser or VS Code.

  • JavaScript
  • VSCode Extension
  • Safetensors
  • GGUF
  • Hugging Face
More details & ways to try it

A tool for exploring ML model architectures from checkpoint metadata. It shows how layers connect and estimates memory requirements without loading the full weights. Available in the browser and as a VS Code extension.

WhatsApp · text & voiceMaurice
GmailRead unread emailGoogle CalendarRead events · scheduling plannedNotionSave notes & tasks · plannedGranolaUse meeting notes · planned

Maurice

An open-source personal assistant you can text or speak to in WhatsApp. The prototype checks connected Gmail inboxes and Google Calendar. The goal is to save notes and tasks in Notion, pull action items from Granola meeting notes, schedule tasks on request, and send one briefing with email, events, and tasks.

  • Python
  • WhatsApp
  • Hermes
Deep learning · from scratch

Deep Learning Libraries (from scratch)

NumPy implementations of neural networks, from multilayer perceptrons to transformers, with forward and backward passes built from scratch.

  • Python
  • NumPy
  • SciPy
  • Cython
  • PyTorch
  • PyPI
Explore the collection · 4 projects

A collection of NumPy implementations covering multilayer perceptrons, convolutional networks, recurrent networks and transformers. I implemented the forward and backward passes to understand how the pieces work, then packaged the libraries so others could use them.

  • npmlp-core

    Modular MLP framework with linear layers, six activations (ReLU, Sigmoid, Tanh, GELU, Swish, Softmax), Batch Normalization, and vectorized optimizers.

  • custom-cnn

    CNNs from scratch: Conv1d/2d, transposed convolution, max/mean pooling, up/down-sampling, and 7 activation functions.

  • custom-rnn

    RNN & GRU cells with full BPTT, the CTC forward-backward algorithm, CTC loss, and greedy/beam-search decoders.

  • custom-transformer

    Multi-head attention with hand-derived gradients; a Pre-LN encoder-decoder supporting CTC + cross-entropy ASR and decoder-only language modeling.

Agentic systems

Phoenix: Agentic Software Engineering

AI agents that investigate GitHub issues, propose changes and test fixes, with controls for human review.

  • Python
  • Multi-Agent LLMs
  • LangChain
  • FastAPI
  • GitHub API
  • SWE-bench
Explore the collection · 3 projects

I co-developed Phoenix to investigate GitHub issues, propose changes and test fixes, with controls for human review. The research was accepted at IEEE IRAI 2026. It brings together my interests in language models, developer tools and building systems whose decisions can be checked.

  • Phoenix: Safe GitHub Issue Resolution (IEEE IRAI 2026)

    Accepted at IEEE IRAI 2026. A multi-agent system for investigating GitHub issues, writing and testing changes, and preparing pull requests for human review.

  • Phoenix Agent

    Autonomous code-analysis agent running a 7-phase control loop with human-in-the-loop approval, parallel CoderAgents, WebSocket streaming, and a 3-layer memory system (Redis + PostgreSQL + Neo4j).

  • Phoenix RAG

    Retrieval-Augmented Generation for code refactoring with a ReAct reasoning agent, ChromaDB semantic search, and groundedness verification to reduce hallucinations.

Research & publications

Research across AI agents, brain-computer interfaces and engineering systems, including published papers, accepted work and preprints.

Read more

VLAKit

A toolkit for distributed, resumable VLA fine-tuning, with experiment configuration, checkpoint recovery and model-weight uploads. Built around the practical needs of running experiments on rented GPU machines.

Read more

Experience

Where I’ve made an impact.

Research, engineering, and the people I’ve built with.

Full experience

IEEE Rwanda Sub-Section

Dec 2025 - Present

Current
  1. SAC Chair

    Aug 2026 - Present

    Chairing the Student Activities Committee (SAC) for IEEE Rwanda Sub-Section, supporting student branches and coordinating student engagement and collaboration across universities.

    Role details
    • Leadership
    • Student Activities
    • Community
  2. Treasurer

    Dec 2025 - Present

    Serving as Treasurer of IEEE Rwanda Sub-Section, managing sub-section finances.

    Role details
    • Financial Management
    • Leadership
    • Community

Logic & Matter Labs

May 2023 - Present

Current
  1. Founder

    May 2023 - Present

    Founded an initiative mentoring K-12 students in web development through hands-on workshops at schools.

    Role details

    Founder

    • Education
    • Mentorship
    • Web Dev

Neotix Robotics

Jun 2026 - Oct 2026

  1. Member of Technical Staff - VLA

    Sep 2026 - Oct 2026

    I worked on robot-learning experiments and the systems around them: simulation tasks, model fine-tuning, policy evaluation and physical deployment.

    Role details
    • VLAs / VLMs
    • Robotics
    • Reinforcement Learning
    • LeRobot
  2. Research Intern - VLA

    Jun 2026 - Aug 2026

    Fine-tuned MolmoAct 2 and π0.5 on teleoperation data and deployed inference on bimanual YAM robots.

    Role details
    • VLAs / VLMs
    • Robotics
    • LeRobot
    • Hugging Face

Carnegie Mellon University Africa

Sep 2024 - May 2026

  1. Graduate Teaching Assistant

    Sep 2024 - May 2026

    Supported courses in deep learning, computational materials science and the Bridge Program, helping students work through programming, models and experiments.

    Role details

    Smart Africa Scholar

    • Deep Learning
    • PyTorch
    • Materials Science
  2. Graduate Research Assistant

    May 2025 - Aug 2025

    Used computational analysis and Python experiments to study relationships between material structures and experimental outcomes.

    Role details

    Summer Research

    • Research
    • Python
    • Materials Science

Research & publications

Research across AI and engineering.

All publications
2026Preprint

Zero-Shot Neural Priors for Generalizable Cross-Subject and Cross-Task EEG Decoding

arXiv preprint · Signal Processing (eess.SP)

Baimam Boukar Jean Jacques, Brandone Fonya, Nchofon Tagha Ghogomu, Pauline Nyaboe, Kipngeno Koech

Abstract

A zero-shot cross-subject framework for generalizable EEG decoding on the large-scale Healthy Brain Network dataset, benchmarking a CNN baseline, a hybrid LSTM, and a Transformer-based foundation model. To adapt the Transformer for regression without catastrophic forgetting, we propose a novel progressive unfreezing strategy. The fine-tuned Transformer reaches an nRMSE of 0.9799 on unseen subjects (vs. 0.9991 for the baseline), advancing scalable, calibration-free EEG decoding for computational psychiatry and behavioral prediction.

  • Brain-Computer Interfaces
  • EEG
  • Foundation Models
2026Accepted

Phoenix: Safe GitHub Issue Resolution via Multi-Agent LLMs

IEEE International Conference on Responsible Artificial Intelligence (IRAI), Melbourne, Australia

Kipngeno Koech, Muhammad Adam, Baimam Boukar Jean Jacques, Joao Barros

Abstract

A multi-agent LLM system that resolves GitHub issues from triage through pull-request creation, combining seven layered safety controls with a baseline-aware test evaluation strategy. Work is decomposed across six specialized agents (planner, reproducer, coder, tester, failure analyst, and PR agent) coordinated by a label-based GitHub webhook state machine, with every change checked against a baseline test run before a PR is opened. Phoenix oracle-resolves 75% of a SWE-bench Lite slice with no pass-to-pass regressions, and preserves correctness on 100% of a 42-issue pilot across 14 repositories.

  • Multi-Agent LLMs
  • Software Engineering
  • SWE-bench
2025Published

Improving SSVEP BCI Spellers with Data Augmentation and Language Models

IEEE Workshop on Signal Processing Systems (SiPS), Hong Kong

J. Zhang, R. Zhang, K. Koech, D. Hill, and K. Shapovalenko

Abstract

A hybrid framework integrating domain-specific EEG data augmentation with a language model to improve SSVEP-based Brain-Computer Interface speller accuracy for individuals with motor disabilities. It decodes scalp-recorded EEG to identify the characters a user gazes at, tackling high EEG variability and poor generalization to unseen subjects with an end-to-end PyTorch training and evaluation pipeline.

  • Brain-Computer Interfaces
  • EEG
  • NLP

The person behind the work

Models, experiments and software.

I’m an AI engineer and researcher based in Kigali, Rwanda. I completed my M.S. in Engineering Artificial Intelligence at Carnegie Mellon University in 2026.

I’ve built tools for inspecting large ML models, developed systems where AI agents collaborate on code, and worked on brain-computer interface research. At Neotix Robotics, I trained models and helped get them running on physical robots.

I enjoy tracing a training failure, working through the math behind a model, or turning an idea into a tool someone else can run. I also teach, contribute to open source, and support student activities through IEEE Rwanda.

Outside work, I spend time hiking, travelling and reading.

Away from the keyboard

Education

A foundation for the work.

Jul 2024 - May 2026

Master of Science in Engineering Artificial Intelligence

Carnegie Mellon University

Kigali, Rwanda

Details & coursework

QPA: 3.77/4.00. Graduate teaching assistant and researcher, with work in deep learning, brain-computer interfaces and computational materials science.

  • Computer Vision
  • Deep Learning
  • Probabilistic Graphical Models
  • Stochastic Processes
  • Mathematical Foundations of Machine Learning
Sep 2020 - Nov 2024

Bachelor of Science - Software Engineering

Multimedia University of Kenya

Nairobi, Kenya

Details & coursework

Best Club of the Year 2022. Class Representative throughout. Vice Chair of MMU Tech Community. Foundation in software engineering and algorithms.

  • Software Engineering
  • Algorithms
  • Data Structures
  • Systems Design
Jun 2022 - May 2023

Software Engineering Bootcamp

Holberton School (ALX Africa)

Remote

Details & coursework

Intensive software engineering program covering full-stack development, system design, and professional engineering practices.

  • C Programming
  • Python
  • DevOps
  • Web Development

Away from the code

Notes on being human.

Essays, reflections, and things I’m thinking about.

Research, engineering and collaboration

Have something
in mind?

I’m open to research collaborations and engineering opportunities. I’m also exploring PhD opportunities for 2027 in machine learning, AI and robotics.

Get in touchkip@kipngenokoech.com