albertbustamante.com
This site — a hand-built static site with a custom terminal-inspired theme, dark mode, a hero carousel, and a growing Tech & Living blog.
HTML
Bootstrap 5
JavaScript
About Me
Tech entrepreneur focused on AI automation, new technology based products, and developer tooling.
Bio
I have spent over two decades building software and systems—from data pipelines and APIs to web UX. I have a passion for engineering, art, and traditional skills. My recent work varies, often blending engineering and art to solve real-world problems and enhance daily life.
When I’m not working on software or work projects, I’m probably cycling, tweaking a sourdough starter, or planning the next garden season.
Portfolio
A selection of projects and experiments — current work up top, a few older builds and exercises below. More lives on GitHub.
Current work
Site AI Assistant
A Claude-powered assistant that answers visitor questions grounded in the site's own content, served through a lightweight server-side proxy with prompt caching and rate limiting.
PHP
Claude API
JavaScript
Code & experiments
Landmarks
A SwiftUI app built while exploring Apple's declarative UI framework and the iOS development workflow.
Swift
SwiftUI
adventure_game
A text-based adventure game written in Python — a small exercise in game loops, state, and parsing player input.
Python
Harvard CS50 — Web
Coursework from Harvard's CS50 Web Programming with Python and JavaScript (edX) — front-end through full-stack fundamentals.
HTML
JavaScript
Python
Recent Training
A snapshot of recent, hands-on coursework — mostly DeepLearning.AI short courses — focused on building with modern generative AI. Each was project-based; the notes below are what I built or worked with.
Agents, Protocols & Tooling
Claude Code: A Highly Agentic Coding Assistant
Agentic coding workflows — planning, subagents, MCP integration, git worktrees, hooks, and automated GitHub PR review.
AnthropicMCP: Build Rich-Context AI Apps
Built and deployed Model Context Protocol servers and clients to standardize how LLMs reach tools, data, and prompts.
AnthropicLong-Term Agentic Memory with LangGraph
Built an email agent with semantic, episodic, and procedural long-term memory using LangGraph.
LangChainACP: Agent Communication Protocol
Cross-framework agent communication; built sequential and hierarchical multi-agent systems (since merged into A2A).
IBM ResearchBuilding Code Agents with smolagents
Code agents that act by writing executable code; safe sandboxed execution and production-grade evaluation.
Hugging FaceOpen & Multimodal Models
Building with Llama 4
Long-context (up to 10M tokens) and multimodal apps — image grounding, reasoning, prompt optimization, and synthetic data.
MetaOpen Source Models with Hugging Face
NLP, audio, vision, and multimodal tasks with the transformers library, deployed via Gradio and Spaces.
Hugging FaceBuilding Generative AI Applications with Gradio
Fast GenAI app UIs — summarization, image captioning, diffusion image generation, and LLM chat.
Hugging FacePrompting & Python Foundations
ChatGPT Prompt Engineering for Developers
Systematic prompting for summarizing, inferring, and transforming text, plus building a custom chatbot on the OpenAI API.
OpenAI · DeepLearning.AIPrompt Engineering with Llama 2 & 3
Few-shot and chain-of-thought prompting, Code Llama pair-programming, and Llama Guard safety checks.
MetaAI Python for Beginners
Practical Python that calls LLMs, automates tasks, and analyzes data with BeautifulSoup, matplotlib, and live APIs.
DeepLearning.AIEvaluation & MLOps
Evaluating and Debugging Generative AI
Experiment tracking, dataset and model versioning, and tracing LLM prompts and responses with Weights & Biases.
Weights & Biases