Member of Technical Staff Intern

Why this work matters

Touchdown starts with people doing real work: teach the team, build one useful AI system, manage what runs, and measure what still breaks. We follow repeated limits through software, inference, kernels, memory, hardware, materials, and manufacturing research only when the evidence earns the next layer. The ambition is broad. The work is always bounded by a user, a test, and a receipt.

Overview

This is a broad technical internship, not a front-end-only internship and not a passive shadowing program. You will be assigned one primary technical lane and one bounded project: agent and AI systems, product and backend software, inference infrastructure, profiling and kernel tooling, FPGA or driver test infrastructure, hardware or materials research tooling, manufacturing-data analysis, or interactive technical education. Every assignment has one named owner, one bounded deliverable, one acceptance test, and one review path. It will also name the user, interfaces, methods, expected artifact, review cadence, and handoff. Breadth means you learn how your work connects to the system; it does not mean doing unowned miscellaneous tasks.

What you will own

• Build one bounded project in an assigned lane across agent development, product software, backend systems, inference, developer tooling, kernels, FPGA or drivers, hardware systems, materials or manufacturing data, research tooling, or education.
• Start from a written problem, user, expected result, interfaces, constraints, and acceptance test.
• Use AI tools to learn, prototype, debug, and move faster while keeping responsibility for testing and correctness.
• Read real repositories, documentation, traces, papers, specifications, experiment records, or process data instead of stopping at generated summaries.
• Trace code and system behavior through evidence and label source-backed, fixture-backed, simulated, and live results correctly.
• Write a clear handoff that explains what changed, how to run it, what the evidence proves, and what remains unproven.

Technical territory

Projects may include agent tool loops, context and evaluation, web and backend systems, APIs and data, observability, model serving, routing and caching, workload replay, profiling, compilers and GPU kernels, FPGA and driver test infrastructure, simulation and verification, architecture models, experiment automation, DOE or yield-data analysis, metrology and reliability data, documentation, or interactive technical education. The project is selected based on current work, your demonstrated ability, supervision, safety and access constraints, and a scope that can be completed and reviewed.

Representative outputs

• A written project brief naming the user, problem, technical lane, interfaces, constraints, acceptance test, evidence state, reviewer, and delivery date.
• A working artifact such as code, tests, a benchmark, simulator, data pipeline, experiment tool, verification harness, visualization, or interactive lesson.
• A review packet with source links, commands or methods, results, failures, limitations, and the evidence that supports each claim.
• A final demonstration and handoff that another engineer, researcher, or educator can reproduce and extend.

What success looks like

• You ship one useful artifact that passes its defined acceptance test.
• You can explain the full path from user problem to code, system behavior, evidence, and limitation.
• Your tests and documentation let another person reproduce and extend the work.
• You leave with deeper technical judgment, not only familiarity with one tool or model.

What you bring

• Current bachelor’s, master’s, or PhD student, or recent graduate, with strong fundamentals and evidence of building, experimenting, or researching.
• Ability to learn unfamiliar tools, ask precise questions, and finish a bounded project.
• Code, research, experiments, technical writing, or projects that demonstrate curiosity and ownership.
• Comfort reading existing code and documentation before proposing a rewrite.
• Basic testing and debugging discipline appropriate to your area.
• Clear written communication and willingness to receive direct technical review.

Helpful experience

• Agent systems, evaluation, or developer tools
• Full-stack, backend, data, or infrastructure engineering
• ML serving, routing, caching, profiling, or inference infrastructure
• Compilers, GPU kernels, FPGA, drivers, architecture, or verification
• Materials science, semiconductor processing, packaging, metrology, reliability, DOE, or manufacturing data
• Research communication or technical education

How the role works

• Internship.
• San Francisco / Bay Area preferred. Remote within the United States may be considered for the right person.
• Scope, start date, and employment details are discussed during the process.