Hi 👋, I'm Jake McIntosh

Self-Taught AI Ops Tinkerer · Local-First Agent Systems Builder · Data Pipeline Orchestrator

---

🧭 About Me

I build local-first AI systems that connect agents, tools, files, memory stores, APIs, queues, and data pipelines into working backend architectures.

I am entirely self-taught and especially interested in how information moves through a system:

Files / APIs / Events
        ↓
Ingestion + Cleaning
        ↓
Queue / State / Workers
        ↓
Embedding / Search / Simulation
        ↓
Memory / Validation / Tool Calls
        ↓
Agent Output / API Response / Human Review

I am currently looking for an entry-level Data Ops, Junior Python Developer, or AI/Automation Engineering role where I can bring practical build experience, persistence, and systems thinking into a structured team.

---

🎯 What I’m Looking For

I can bring Strong self-directed learning Practical Python project experience Local AI and RAG experimentation Backend orchestration mindset Persistence with messy systems Documentation-heavy thinking Willingness to learn from code review

I want to grow in Production Python structure CI/CD and GitHub Actions Testing with pytest Docker and deployment hygiene Clean API design Code review discipline Maintainable syntax and patterns

---

🛠️ Tech Stack

Core Languages

AI, Data, and Search

Backend, Tools, and Infrastructure

---

🧠 Architecture Mindset

I focus on getting large, theoretical systems to run end-to-end first.

That usually means proving the full flow works, then improving the structure:

Prototype → Working Pipeline → State Tracking → Validation → Documentation → Refactor → Tests

I care about:

• data movement across tools and services;
• agent/tool boundaries;
• semantic memory and retrieval;
• worker orchestration;
• local-first development;
• debuggable systems;
• clear documentation for future maintainers.

I am still learning production polish, but I am comfortable tackling complex architectures and breaking them into inspectable parts.

---

📌 Featured Projects

Project	What it Shows	Stack / Concepts
Agent Backend	Local agent backend with daemon logic, MCP/API gateway patterns, JSON-RPC style flow, runtime state, and semantic memory ideas	Python Node.js SQLite JSON-RPC MCP RAG
Quantule Mapper	Physics/search bench with worker orchestration, API notes, notebooks, and simulation workflow structure	Python FastAPI Workers WebSockets Simulation
ToolSet	Developer-agent utilities for large repositories, context extraction, file maps, and semantic slicing	Python Repo Mapping Context Tools
Alethiea RAG System Legacy	Local RAG architecture exploring canonical recall, semantic retrieval, and LM Studio-style workflows	Python RAG ChromaDB Local LLMs
IRER Test Bench	Experimental physics/simulation framework with solver, validation, and service-style architecture	Python JAX FastAPI Docker React

---

🧪 Project Deep Dives

🤖 Agent Backend — Local Agent Orchestration

A local-first backend system exploring how agents can call tools, access memory, manage runtime state, and interact with local model workflows.

Highlights

• Python daemon-style backend
• MCP/API gateway thinking
• JSON-RPC style communication
• SQLite-backed runtime state
• Semantic memory and retrieval concepts
• LM Studio-compatible local LLM experimentation

Why it matters

This is the project that best represents my long-term direction: backend infrastructure for local AI systems.

🧬 Quantule Mapper — Physics Bench + Worker Search

A simulation/search environment focused on orchestrating workers, experiments, APIs, notebooks, and local runtime commands.

Highlights

• Worker daemon structure
• API/WebSocket notes
• Notebook walkthroughs
• Runtime command documentation
• Simulation/search experimentation

Why it matters

This project shows my interest in long-running systems, search loops, telemetry, and experimental infrastructure.

🧰 ToolSet — Developer-Agent Workflow Utilities

A collection of utilities for reducing friction when using AI agents and cloud coding platforms with large repositories.

Highlights

• File mapping
• Semantic slicing
• Repository context extraction
• Developer-agent workflow helpers
• Large-file and large-repo navigation support

Why it matters

This project shows how I think about tooling around the developer experience, not just the application layer.

---

📊 GitHub Stats

---

🧰 Tools and Technologies

Python	C++	JavaScript	TypeScript	FastAPI	SQLite
Docker	Git	GitHub	VS Code	React	Linux

---

🧭 Current Learning Roadmap

flowchart LR
    A[Working Local Systems] --> B[Cleaner Python Packages]
    B --> C[Tests with pytest]
    C --> D[GitHub Actions CI]
    D --> E[Dockerised Services]
    E --> F[Junior Production Engineering]

Loading

I am currently focusing on:

• cleaner Python module boundaries;
• pytest test coverage;
• GitHub Actions;
• Docker-based reproducibility;
• API documentation;
• better README structure;
• smaller, more reviewable commits.

---

🧩 Open Source Philosophy

My public repositories are designed to show how I think.

I keep core engines, toolchains, context extractors, mapping utilities, and experimental systems open where possible. Bespoke workflows, private agent configurations, and commercially sensitive structures stay private to respect security and IP boundaries.

The public work is meant to demonstrate:

• architecture thinking;
• persistence;
• documentation habits;
• willingness to experiment;
• ability to stitch systems together;
• areas where mentorship can sharpen the craft.

---

🤝 Let’s Connect

I’m based in Preston, UK and actively looking for my first professional technical role.

I’m especially interested in Data Ops, Junior Python Development, AI Ops, backend automation, and RAG/agent infrastructure.

---

“Building local AI systems, one daemon, vector store, queue, and broken test at a time.”