# Brayan Njoroge

Brayan Njoroge is a remote worker on Himalayas, OS Research Fellow at Manifold Research Group, based in Kenya.

Source: https://himalayas.app/@brayannjoroge

```text
AI / Machine Learning Engineer at Manifold Research Group, contributing to the MultiNet open-source AI-agent benchmark.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | Kenya |
| Authorized to work in | Kenya |
| Salary expectations | 120k+ USD |
| Job categories | Machine Learning Engineer, AI Engineer, Data Engineer, Research Engineer |
| Skills | LLM Post Training, Reinforcement Learning from Human Feedback (RLHF), Policy Distillation, Model Evaluation, Model Evaluation Harnesses, LLM Training, Incentive Modeling, SFT (Supervised Fine Tuning), Proximal Policy Optimization (PPO), Group Relative Policy Optimization (GRPO), Direct Preference Optimization (DPO), Natural Language Processing, ML Deployment, Data Pipelines, Confidence Intervals |
| Website | https://brayanbrayan.github.io |
| GitHub | https://github.com/Brayanbrayan |

## About

```text
At Manifold Research Group, I contribute to MultiNet v2.0, an open-source benchmark for AI agents. I added full-context history and a per-turn step budget to the agent runner, and designed prompt-caching support for Anthropic, Kimi, and OpenAI models.

On a project reproducing on-policy distillation, I built a GSM8K evaluation harness and compared the student model with its teacher. I’m running the work on a Colab TPU v5e chip with Tunix.

I built a GPT-style language model from scratch in PyTorch, then implemented PPO, GRPO, and DPO for an RLHF alignment study. The experiments revealed reward-margin explosion in DPO and group collapse in GRPO; targeted tuning reversed the ranking and gave GRPO +3.43 average reward.

My other projects include a modular VLM evaluation pipeline for Qwen3-VL-2B-Instruct and an end-to-end cuisine classifier deployed through a Flask app. I’m an early-career machine learning engineer focused on LLM post-training, evaluation, and end-to-end ML pipelines.
```

## Experience

### OS Research Fellow (current)

Manifold Research Group. Jan 2026 - Present

```text
Contributed to MultiNet v2.0, an open-source benchmark for AI agents, with 3 merged pull requests and 34 commits. Added full-context-window history and per-turn step budget to the agent runner, and designed prompt-caching support for Anthropic, Kimi, and OpenAI models.
```

Skills: AI-Agents, Benchmarking, Prompt-Caching, PR-Workflow, Code-Review, Python

## Education

### KCA University

Bachelor of Science, Information Technology

```text
Pursuing a Bachelor of Science in Information Technology at KCA University in Nairobi.
```

## Tech stack

- GitHub
- Jupyter
- NumPy
- Pandas
- JavaScript
- Python
- C++
- TensorFlow
- PyTorch
- scikit-learn
- NLTK
- Gradio
- Flask
- Linux
- Git
- SQL
- Google Colab
- JAX
- Bash
- Grain
- uv
- Matplotlib

To contact Brayan, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Keryo User](https://himalayas.app/@keryouser): Freelance AI Developer, Self Employed
- [Leonard Kipchumba](https://himalayas.app/@leonardkipchumba): AI / ML Engineer, Freelance
- [Lawre Ski](https://himalayas.app/@lawreski): AI Trainer, Outlier
- [Steve M](https://himalayas.app/@stevem): AI Analyst, Turing
- [Kennedy Muriuki](https://himalayas.app/@kennedymuriuki): AI Research Associate, Scale AI
- [Ronald Kibet](https://himalayas.app/@ronaldkibet): AI Engineer, Andela
- [Michael Nganga](https://himalayas.app/@michaelnganga1): Open Source Contributor, Personal Projects
- [Daniel ian Otieno](https://himalayas.app/@danielianotieno): Data Analyst & Quantitative Modeler, Unknown
- [BBL DRIZZY](https://himalayas.app/@bbldrizzy): AI Researcher, Nairobi
- [David Polycarp](https://himalayas.app/@davidpolycarp): Data Science Intern, Africa Center for Data Science and Analytics
