Are you Lucid?

The Role

We're looking for a hardware engineer to own the silicon-level foundations of our technology. You'll bridge the gap between hardware security features and AI workloads — building verifiable computation pipelines where we can statically derive what a GPU kernel should compute and cryptographically prove that it did. You'll work directly with silicon vendors, design attestation systems for modern accelerators, and build the tamper-proof link between a model's architecture and its execution on hardware.

What You'll Do

• Design verifiable compute pipelines that derive hardware behavior expectations from model definitions and tiling configurations, then prove execution integrity
• Build low-level software (C, C++, Rust) that interfaces with TEEs, secure enclaves, and cryptographic accelerators to attest to chip location and computation integrity
• Implement static kernel analysis tools to extract memory access patterns and computational structure from GPU kernels (Triton, CUTLASS, PTX/SASS)
• Serve as technical liaison with silicon vendors (NVIDIA, AMD, Intel) on security architectures, side-channel implications, and next-generation accelerator roadmaps

What You'll Bring

• BS, MS, or PhD in Computer Engineering, CS, or EE with a focus on computer architecture, compilers, or systems security
• Deep understanding of GPU architecture — memory hierarchies, tensor core operations, tiling strategies, and how logical tensors map to physical hardware
• Strong systems programming in C, C++, or Rust with experience interfacing with hardware security primitives
• Experience with CUDA, Triton, or CUTLASS kernel programming

Bonus

• Familiarity with MLIR/LLVM or compiled binary analysis
• Experience with hardware performance counters (Nsight Compute, CUPTI)
• Knowledge of TEEs (Intel SGX/TDX, AMD SEV)
• Interest in formal methods or layout algebra

We are open to hiring both early-career and experienced candidates. For the latter, we are open to higher compensation and more senior titles. We will also consider exceptional candidates on a part-time basis.

The Role

You'll own our prosumer product experience end-to-end — the product that a technically curious person installs, connects to their AI tools, and actually uses every day.

This is not a "build to spec" role. You'll take high-level outcomes ("a user should be able to chat with their agent and see what it did") and drive them from prototype to alpha to beta to GA, making deliberate scope and quality tradeoffs at each stage. You'll decide what to ship now, what to cut, and what to defer — then ship it, put it in front of users, and iterate.

You'll work closely with the founding team on product direction and with our architecture lead on the underlying platform. You'll inherit a large, architecturally sound monorepo that has accumulated complexity — your job is to navigate it, simplify where needed, and ship a working product without needing to rewrite everything.

What You'll Do

• Own the prosumer product experience: chat interface, tool connections (MCP), activity timeline, approvals workflow, onboarding — the full surface a user interacts with
• Drive features from prototype to production: take modules from rough monorepo state to alpha-ready, then iterate toward GA based on real user feedback
• Make product decisions: scope features, cut what doesn't matter, prioritize what does — you have product judgement, not just engineering skills
• Ship fast using AI tools: you already use AI-assisted development (Claude, Cursor, Copilot, or similar) as a core part of your workflow and understand the quality tradeoffs — when to accept AI output, when to rewrite, when it's saving you days vs. creating debt
• Run build-test-learn loops: get features in front of users quickly, gather signal, adjust — not waterfall, not endless polish, just fast informed iteration
• Work autonomously over long sessions: take a goal, break it down, execute across multiple files/systems, course-correct without waiting for direction

What You'll Bring

• 3+ years full-stack engineering, ideally with SvelteKit/Svelte and Python (FastAPI). Strong in modern frontend — you can build polished, functional interfaces, not just APIs.
• Demonstrable AI-leveraged development workflow: show us how you use AI tools to move faster. We don't care which tools — we care that you've integrated them deeply and understand when they help vs. when they hurt.
• Product instinct: you look at a feature and see "the MVP is this, cut that, defer this other thing" without needing a product manager to tell you. You've shipped things that users actually use.
• Comfort with complexity: you can navigate a large monorepo, understand dependency graphs, work with feature flags, and ship incrementally without needing everything to be clean first.
• Hands-on DevOps mindset: you're comfortable owning your code through deployment — working with containers, CI/CD pipelines, and cloud environments. Experience with Kubernetes (or similar orchestration systems) is a strong plus, especially for shipping and operating distributed services.
• Self-directed and high-agency: given "make this module alpha-ready," you figure out what that means, what's blocking it, and how to get there. You ask good questions when stuck, not permission to proceed.
• Strong communicator: you document decisions, flag risks, and write clearly. Your documentation becomes the onboarding material for the next person.

Nice to Have

• Experience with the MCP (Model Context Protocol) ecosystem — servers, clients, tool federation
• Familiarity with AI agent architectures, LLM application patterns, or conversational interfaces
• Experience shipping a prosumer or PLG product from 0 → 1
• Understanding of AI compliance and governance (EU AI Act, ISO 42001, NIST AI RMF) — relevant as we expand to enterprise surfaces
• Background in security infrastructure, trusted execution environments, or cryptographic audit systems — Lucid's deeper technical stack includes these and you'd eventually touch them
• Monorepo tooling experience (turborepo, shared packages, workspace management)

Why This Matters

AI agents are getting powerful fast. The tooling for making sure they're trustworthy, controllable, and transparent is lagging behind. We're building the product that closes that gap — starting with individuals who want to use AI agents confidently, and expanding to teams and organizations that need governance at scale.

You'd be building something you'd use yourself: an AI assistant that's actually transparent about what it does, with guardrails that protect you without getting in the way.

What We Offer

• Founding-stage equity in an AI infrastructure company at the right moment
• Direct collaboration with the founding team — your input shapes the product, not just the code
• A codebase built on solid architectural primitives (not a throwaway prototype)
• The chance to build the trust layer for AI agents before anyone else does

How to Apply

Send us:

• A link to something you shipped (product, project, open-source contribution)
• A short note on how you use AI tools in your development workflow
• Optionally: your take on what "trustworthy AI" means to you as a builder