Stanford EE student · Class of 2029

Diego
Sanchez Tirado.

I design circuits and build software.

I'm a Stanford undergraduate, class of 2029. I research memory systems for AI inference and build hardware and software.

About my work

Stanford EE · Class of 2029

Design, build,
test, repeat.

I design custom circuit boards, model computer architectures, and develop software. I test what I build on real hardware and in working applications.

At Stanford, I'm studying electrical engineering, class of 2029. My interests span computer architecture, robotics, and tools that make everyday work better.

Hardware & research

Model memory.
Build hardware.

I model memory architectures for AI inference with Stanford’s Engineering Robust Systems Group. My independent projects include custom circuit boards and embedded hardware.

Stanford Robust Systems Group · Independent builds

Software & developer tools

Build tools.
Verify software.

For Stanford's Modern Software Developer course, I'm exploring coding agents, MCP, and how to verify the software people actually use.

Applied AI / Verdnt Group

Develop AI
applications.

At Verdnt Group, I build AI systems that turn scattered information into useful tools. My work includes beverage intelligence, hospitality prototypes, and a plate-scanning workflow.

Contact

Work with me.

Contact me about hardware, software, or research opportunities.

Experience & education

A few stops
along the way.

2026 - present

Verdnt Group

Building AI systems for businesses, including FStea's beverage intelligence engine.

2026 - present

Stanford Robust Systems Group

Undergraduate research in area-scalable memory for LLM inference.

Summer 2026

Amazon Lab126

Hardware Development Engineer Intern and Robotics Apprentice. Engineering knowledge systems, retrieval safeguards, and bimanual robot manipulation.

2025

Code Four (YC X25)

GTM Team Lead. Built outreach automation and helped onboard 520+ officers.

2023 - 2025

Duke University

Research across nanomaterials, thin films, and cyber threat intelligence.

Expected May 2029

Stanford University

B.S. Electrical Engineering

Read my full resume ↗

Research & independent builds

From the model
to the workbench.

Stanford Robust Systems Group

Rethinking memory for AI inference

With Prof. Subhasish Mitra and Sam Dayo, I model how memory capacity, bandwidth, and network topology shape LLM inference, comparing area-scalable mesh memory with HBM.

Bowman Brockman Grant

ErgoTouch

A custom split keyboard, from a resin-printed case and KiCad PCB to hand-soldered switches and RP2040 firmware.

Independent build / 2026

A bike, rebuilt

A mountain-bike conversion with a custom 48 V battery pack and mid-drive motor. Built, tested, and put to work on my Amazon commute.

Digital logic & energy harvesting

Logic you can touch

A gate-level 4-bit adder/subtractor with a two-digit display, plus a piezoelectric rig that harvests typing force.

CS146S / Fall 2026

Notes from
the developer's bench.

  1. Week 1

    Inside the agent loop

    Inspecting a coding-agent session to understand prompts, tool calls, and context.

  2. Week 2

    From plate to persistent log

    A five-tool MCP server connecting Drive photos, plate analysis, and a Sheets meal log, with OAuth and a preview before writing.

  3. Week 3

    Check the thing people use

    A reusable skill for checking FStea AI's published interface, mobile layouts, languages, and saved results.

CMUX and Pi links point to the upstream tools. Course submissions are private.

Explore the course ↗

Selected work at Verdnt Group

Information
into action.

Beverage intelligence

FStea AI

Nine collectors bring together social, marketplace, retail-launch, search, and seasonal signals. The engine turns them into ranked trends, product concepts, and product specifications.

Hospitality prototypes

A better winery day

Two interactive concepts: guest preparation for Hill Family Estate and a portfolio view for Soda Rock Winery.

Independent demos using fictional data.

Food meets software

Plate Scanner

A phone-first tool for estimating what's on a dining-hall plate. My course work adds an MCP server connecting plate photos in Google Drive to a meal log in Google Sheets.