Web3 is a new paradigm

Sherlock exists to make open finance secure enough for billions of people to trust with real value. We believe sound money and reliable financial infrastructure should be basic human rights and that Web3 can only achieve that if it’s secure by design.

Traditional audits alone don’t solve that problem. Sherlock connects every layer of protocol security through collaborative audits, audit contests, bug bounties, coverage, and AI-powered analysis. Each product reinforces the others, forming a lifecycle system that protects protocols from development through live operations.

Sherlock is the only security provider that offers financial coverage after every audit, backing its work with up to $500K in post-launch exploit protection. The company has run more public audits than anyone in the space, securing code for the Ethereum Foundation, Aave, Optimism, Cosmos, MakerDAO (Sky), and dozens of other top teams.

Sherlock is backed by Alliance, Archetype, IDEO, CoinFund, Lattice, Spartan, A Capital, and other leading Web3 and Silicon Valley investors.

Sherlock's Culture
Sherlock is a performance-driven, high-ownership environment built for people who care about the mission and can handle the pace. The team shares a few common traits:

The Role

Sherlock is building an AI product for blockchain security. The core is already built and in customers' hands. What it now needs is someone to sit between our customers — protocol teams shipping real money on-chain — and the product, and own the loop from "what do they actually need" to "shipped."

This is a generalist product engineering role. You will talk to customers, decide what matters, and ship it yourself. You will move across the entire product surface: the UI that protocol teams use to submit code and read findings, the prompting and judging pipelines that decide which issues reach the customer, the benchmarks that tell us whether the system is actually getting better, and the product calls in between. The role is explicitly broad. We are looking for breadth, judgment, and ownership over deep specialization in a single area.

We are looking for a senior Python developer with a strong track record shipping software products to real users. You ship UI as comfortably as you ship backend — not necessarily as a designer, but you enjoy the work and have done it before. You use LLMs in your daily workflow and have shipped at least one feature whose value depends on LLM output quality. You are entrepreneurial, highly self-directed, and product-minded.

If you're serious about applying for this job and you've read this far, please email product at sherlock dot xyz with 4-sentence answers to each of these three questions: What is the most recent Python-based product you shipped to real users, and what was your role across product decisions, backend, and UI? This role spans customer conversations, UI work, prompting/eval, and product judgment in a single seat — which of these have you done before, and which would be new for you? Why should Sherlock select you over hundreds of other applicants?

Your role is simple:

Own the loop from customer need to shipped product across Sherlock's AI product — UI, prompting, benchmarking, and the product calls in between.

Sherlock already has a rockstar team of developers and security researchers who can shore up your weaknesses in certain areas and ship a useful, production-ready application. We are willing to invest deeply in your success, whatever it may require. What we cannot do is hand you a narrow lane. The job sits between customer needs and implementation, and a lot of it is making good calls with incomplete information.

In terms of skills, deep Python expertise and a track record of shipping software products end-to-end are essential. Owning UI work end-to-end (Next.js / TypeScript, in our stack) — thinking through the UX flow, the possible states, and translating that into a working UI yourself — is essential. If you don't enjoy frontend work, this role is not a fit. Daily use of LLMs and at least one shipped feature whose value depends on LLM output quality is expected. You should bring high agency, customer-to-product instinct, and the willingness to put the customer first even when it means adjusting your original vision.

A strong understanding of crypto, DeFi, or smart contract security is a plus, but it's not needed to get started in this role.

Ultimately, your role is to make the product genuinely useful to the protocol teams using it — closing the gap between what customers actually need and what we ship.

The day-to-day responsibilities of this role include:

- Talking directly to protocol teams using the product, identifying what they actually need, and converting those conversations into concrete product and UI changes (10x weight)
- Shipping those product and UI changes yourself — you write the code, not just the spec (9x weight)
- Improving the prompting, judging, deduplication, and false-positive reduction pipelines that determine which findings reach customers (8x weight)
- Building and maintaining benchmarks so we can tell whether changes actually improve the product (7x weight)
- Making product calls in ambiguous situations and pushing work forward without heavy direction (7x weight)
- Working closely with Sherlock's developers and security researchers to stay aligned on goals and ship the right things (6x weight)
- Staying current with practical advances in AI and LLMs that improve developer velocity and product quality (4x weight)

Requirements

- Senior-level Python experience with a track record of shipping SaaS or software products to production
- Comfort owning UI work end-to-end in a modern frontend stack (we use Next.js / TypeScript): thinking through the UX flow, mapping out the possible states a screen can be in, and translating that into a working UI yourself. Not a designer role, but you genuinely enjoy this part of the job.
- Daily use of LLMs in your workflow, plus hands-on experience shipping a feature whose value depends on LLM output quality — prompting, judging, retrieval, evals, or similar. 
- Comfortable on calls with technical customers — this is a recurring part of the job, not occasional. You enjoy turning a customer conversation into a product change.
- High agency: comfortable working without a PM, designer, or eng manager handing you scoped tickets. You decide what to build, scope it yourself, and ship it.
- Working hours that meaningfully overlap with US or EU timezones

Nice to Haves

- Experience setting up evals, judging pipelines, or benchmarking for LLM systems to measure effectiveness and catch regressions
- Familiarity with prompt engineering, evaluation frameworks, retrieval, and vector databases
- Prior experience in fast-moving, underdefined startup environments
- Familiarity with audits, bug reports, or security workflows
- Familiarity with Solidity, EVM tooling (Foundry, Hardhat), or smart contract security

This role is not a fit if you are:

- A pure backend specialist, pure prompt engineer, or pure UI designer
- Someone who can discuss ideas but rarely ships product themselves
- Dogmatic about one function only and uninterested in working across the stack
- Without Python experience or without genuine interest in UI work — both have been the most common reasons we have passed on otherwise strong candidates

What's in it for you

- Attractive base (payable in fiat or crypto) + material tokens/equity + benefits
- Flexible time-off policy
- Great healthcare
- Multiple offsites each year in places like France, Argentina, Thailand, etc.
- Own a core product area at the company leading the AI shift in blockchain security
- Root access to the decision-making process/criteria in all areas of Sherlock and the ability to work directly with the founders
- Move quickly and get stuff done on a small, elite team that has already made a big impact in the crypto space
- Play a huge role in defining the future of Sherlock and accomplishing the goal of making crypto/DeFi accessible to everyone