Skip to content
View Gachuhi-Tech's full-sized avatar
  • Nairobi, Kenya

Block or report Gachuhi-Tech

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.

Content in all repositories owned by your account will be closed.
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
Gachuhi-Tech/README.md

Hi there, I'm Peter Ruthi πŸ‘‹

I'm an Electronics & Computer Technology student passionate about Embedded Systems, IoT, Smart Energy solutions β€” and, more recently, building and evaluating LLM tool-calling agents.

I'm currently pursuing a Bachelor of Science in Electronics & Computer Technology at Mount Kenya University, building on my Diploma in Electrical & Electronics Engineering.

I enjoy designing systems that combine electronics, embedded firmware, and modern web technologies to solve real-world engineering problems β€” and I'm now applying the same systems mindset to AI agents that call tools reliably.


πŸš€ Current Project β€” ESP32-based Smart Energy Monitor

A real-time embedded system designed to monitor electrical parameters and protect connected loads.

Key Features

  • ⚑ Real-time voltage and current monitoring
  • πŸ“Š Power, energy, and power factor measurement
  • 🌐 Built-in web dashboard for live monitoring
  • πŸ”’ Automatic protection against overvoltage, undervoltage, and overcurrent
  • πŸ“‘ IoT-ready architecture for future cloud integration

🧠 Featured Project β€” LLM Agent Completion Portfolio

A hand-built suite of multi-turn tool-use conversations for an LLM assistant, plus the schemas, validator, and a working agent that runs each scenario against a real model.

What's in it

  • πŸ“ 7 hand-written conversation logs covering chained tool use, ambiguity handling, infeasibility, tool failure, and disambiguation
  • 🧩 6 strict OpenAI-compatible function schemas (calendar, email, maps, contacts) with additionalProperties: false
  • βœ… Python validator that checks every conversation against the schemas using jsonschema
  • πŸ€– Groq-backed agent that replays each scenario turn-by-turn against openai/gpt-oss-120b and reports conformance drift
  • πŸ“ An 8-rule quality playbook each conversation is written and reviewed against

Repo: agent-completion-portfolio


πŸ› οΈ Technologies & Tools

Embedded Systems

  • ESP32
  • Arduino
  • Embedded C / C++

Electronics

  • Circuit Design
  • Proteus
  • Electronic Troubleshooting

IoT & Networking

  • HTTP Server
  • Wi-Fi

AI / LLM Engineering

  • Function calling / tool use
  • Prompt engineering & system-prompt design
  • OpenAI-compatible APIs (Groq)
  • Evaluation rubric design

Python & Data

  • Python
  • JSON Schema (jsonschema)
  • REST APIs

Web Technologies

  • HTML
  • CSS
  • JavaScript

Development Tools

  • Git
  • GitHub
  • Ubuntu Linux

🎯 Engineering Goals

I'm continuously improving my skills in:

  • Embedded Systems Engineering
  • Internet of Things (IoT)
  • Smart Energy Monitoring
  • Embedded Linux
  • PCB Design
  • Real-Time Embedded Software
  • LLM Agent Design & Evaluation

My goal is to build reliable systems β€” hardware or software β€” that solve practical engineering challenges through clean integration of components.

Pinned Loading

  1. agent-completion-portfolio agent-completion-portfolio Public

    Multi-turn LLM tool-use conversations: schemas, validator, and a Groq-backed replay agent

    Python