Skip to content

PRIVATE · AI EXECUTION · Ritual

Core Protocol Engineer

A client engineer for stateful precompiles and parallel execution — the hardest seat Ritual has posted, and the one that only matters if the chain is a real roadmap.

You are deciding whether to join a small research company whose last clearly announced round is years behind the postings, on a product that is still easier to describe than to run in production.

Blockchain Lab analysis, for a candidate deciding whether to apply. We are not the employer. Compensation on this posting: Not disclosed.

Listing age

641 days — confirm the req

Equity

Private. Illiquid.

Priced capital we will stand behind

$25m in Nov 2023

Where

Remote

Status
Open on the employer board — checked 2026-10-06
Board updated
2025-01-03 · 641 days before this check
Location
Remote
Remote eligibility
Remote
Employment
Not disclosed
Compensation
Not disclosed
Experience, as stored
Employer asks for deep Ethereum execution-client experience. Years of experience are not stated as a number in the fields we stored.
Original post date
Not disclosed

Apply on the employer siteThe companyPractice project

This seat

Blockchain Lab analysis

Extensions to Geth or Reth, mechanism the research team hands over, transaction pipelines for mixed workloads, and state-conflict work.

It does not own model training, a token listing, or the developer tutorial. If you want to fine-tune weights, this is the wrong req.

The employer names Geth and Reth specifically, plus EIPs. That is execution-client work. People who have only deployed contracts will be screened out by the posting itself, which asks for deep Ethereum client experience.

Stateful precompiles are the product hinge. A precompile that keeps state is no longer a pure function. Conflicts, reversion, and parallel execution stop being papers and become the bug queue.

The board date is 3 January 2025. Twenty-one months later it was still up. Confirm the client fork is staffed.

The company you would actually join

Blockchain Lab analysis

Public reporting and the employer’s own releases. Not a relationship with Blockchain Lab. Not a credit opinion.

Founded
2023
Base
Distributed. No single headquarters is stated on the roles we store.
Status
Private. No token price, circulating supply, or revenue is treated as a fact in this brief.
People
A January 2025 secondary profile put the team near 30. That number is not current headcount.

The technical bet is specific. A blockchain does not run a large model inside the EVM. Ritual's public design inserts a coprocessor: specialised nodes execute a model, then return an output a contract can accept. Infernet is the oracle-shaped version of that idea. The chain, if and when it is the production system, is where stateful precompiles, parallel execution, and heterogeneous workloads are supposed to live. That is why the open roles are not 'AI engineer' in the application sense. They are client, networking, contract-library, and validator roles.

The corporate fact pattern is thinner than the technical writing. One priced round is public. A second is a database row. Several of the Greenhouse records this desk stores were last updated in 2024 or early 2025 and were still listed on 6 October 2026. A listing that has not moved in over a year can be a real job, a forgotten req, or a role whose team no longer matches the text. Confirm the hiring manager is alive to the req before you prepare.

There is also a founder-specific item a candidate should not learn about in week three. In July 2024, reporting said Polychain was investigating Niraj Pant over an alleged financial relationship with portfolio company Eclipse during his time at the firm. This brief found the allegation in secondary coverage and did not find a public finding, settlement, or exoneration. Ask. It is not a verdict, and it is not optional colour.

Founders

Niraj Pant

Co-founder

Previously an investor at Polychain. Contemporaneous coverage credits him with early involvement around projects including EigenLayer and Solana. He is the public face of the round Archetype led.

Akilesh Potti

Co-founder

Described in the November 2023 funding coverage as a quant and builder who had worked at Palantir. The pairing is investor-plus-operator, not two protocol engineers from the same client team.

What the company sells

Infernet

The first product described in public writing: a way for a contract to request an off-chain model run and a proof, rather than a price feed.

Ritual Chain

The execution layer in the company's own framing: stateful precompiles and heterogeneous compute, including AI workloads, on top of Ethereum-client ideas.

Developer path

Frameworks, local environments, and the accelerator language (Altar, Realm) used to pull application teams toward the chain.

Capital

Priced capital we will stand behind

$25m in Nov 2023

Lead investor's own post.

Later round

Not confirmed

A 2025 Series A appears in Preqin only.

Revenue

Not disclosed

No public accounts.

Last priced valuation

Not disclosed

The $25m round did not publish a post-money figure we will repeat.

  1. 8 Nov 2023

    Seed, announced by the lead

    $25 million

    Archetype posted that it was leading Ritual's $25 million round. Blockworks called it a seed. Named participants in that coverage include Accomplice and Robot Ventures. Preqin also lists dao5, Anagram, Hack VC, and Dialectic. Some later blogs mislabel this round a Series A.

  2. March 2025

    Series A, database-reported

    $25 million, unconfirmed here

    Preqin records a further $25 million with Accel, Avra, Hypersphere, and Polychain as new names, plus returning funds. This dossier did not find a Ritual press release for that round. Treat it as a database entry until the company confirms it.

What should change your mind

  • Runway and round status are not something you can read in a filing. Private company, one confirmed round.
  • Mainnet, token, and fee switch are not facts this dossier will assert. The job text sometimes says mainnet. That is the employer's word. Ask what is actually in production.
  • Stale reqs. The smart-contract posting was last updated on the board on 24 July 2024.
  • The category is crowded and narrative-heavy. A coprocessor that cannot show a paid workload is a research project with a payroll.

Ask them

  • Which round is the company actually operating on, and is the March 2025 Series A closed?
  • What is live: Infernet, a testnet, a mainnet, a token? Who pays for inference today?
  • Why has this specific req not been updated, and who would you report to?
  • What is the public status of the Polychain / Eclipse matter?
  • Is the role worldwide remote, or remote inside a list of countries the posting never names?

Sources

How the first months would actually go

Illustrative preparation

Not the employer’s onboarding plan.

  1. Days 1–30: build and run the client the team actually forks. Not upstream vanity.
  2. Days 31–60: one precompile path with a conflict test and a number for the slowdown.
  3. Days 61–90: an EIP-shaped note the research team can argue with. Writing is part of this job.

A week you can rehearse

  • Separate the state transition from the networking story. They fail differently.
  • Name the assumption the client makes about who may call the precompile.
  • Show one test where the honest path and the adversarial path diverge.

What to take into the room

Illustrative preparation

  • Describe a bug you found by reading client code, not by reading a blog.
  • How do you detect a state access conflict before you execute?
  • What would you refuse to put in a precompile?

Proof you can build before you apply

  • A small state machine with an honest path and one adversarial case. The protocol lab is that shape. Bring the code, not the diagram.

Open State-machine simulator. A small state machine with an honest path and one adversarial case.

Walk away if

  • There is no client fork you are allowed to see, even under a private repo, before you join.
  • The role has been rewritten into 'research' with no path to a release.
  • Compensation is deferred entirely to a token that does not exist, and they will not say so plainly.

What the employer actually wrote

Employer stated. The work is inside Ethereum execution clients: stateful precompiles, parallel execution, and heterogeneous compute on Ritual's chain.

  • Design and implement extensions to EVM clients (Geth/Reth) for supporting stateful precompiles
  • Implement novel mechanism design primitives, collaborating with our research team, into the execution layer
  • Develop and optimize transaction processing pipelines for heterogeneous workloads
  • Research and implement solutions for state access conflicts
  • Contribute to core protocol development and EIP implementations

Source: the employer’s public Greenhouse record, job 4614221007. Open the original. First seen by this desk 2026-10-06. Nothing above the fold is a substitute for that page.

A reading of the work

Blockchain Lab analysis. These lines are a reading of the advertisement. They are not extra hiring requirements.

  • Separate the state transition from the networking story. They fail differently.
  • Name the assumption the client makes about the peer set.
  • Show a test where the honest path and the adversarial path diverge.

Still not disclosed

  • Compensation, unless a figure appears in the header. None of these eighteen postings stated one.
  • Closing date, and the day the role was first published. The board timestamp is an update.
  • Which countries may apply. “Remote” alone is not worldwide eligibility. If the employer wrote “worldwide”, that sentence is theirs — still ask which entity can hire you.
  • Visa sponsorship, on-call, and the employing subsidiary.

Not a recommendation to apply. Not an employer assessment. Figures from public reporting move, and a database is not a filing. How this desk labels evidence. Report a listing.

Ritual · Remote · pay not disclosed

Apply