You write to us.
No form, no account to create. Send a short note about the work you want to own, a resume or a profile, and one thing you shipped end to end. We read all of it.
[email protected]Careers
Build the hiring process you would be happy to be judged by.
We are a small team building evaluation infrastructure for companies, staffing firms, universities and job seekers. There is no job board here. There is a description of the work and a way to start the conversation.
Write to [email protected]. A person replies, not an autoresponder.
- No public job board
- You go through our own product
- A person reads every report
What you would be working on
The surface area is bigger than it looks.
One platform serves four audiences. The work spans live voice interviews, sandboxed code execution, scoring, recruiter operations, campus placement software and an MCP server.
The rounds
Five round modes that all score onto one rubric: resume screening, AI interview, human-led interview, AI prompt assessment and coding test. Inside an AI interview there are eight types, from HR Round to System Design. The coding test runs real code against hidden test cases. Making all of it comparable is the hard part.
The report
One share-ready reportcard per candidate: a score out of 100, a verdict line, a skill breakdown from transcript and code, the recording, a searchable speaker-separated transcript, and integrity status. A hiring manager should read it in ninety seconds and still be able to defend it six months later.
The workspace
A kanban ATS pipeline, multi-round workflows, Clients, the workspace layer that lets agencies run many accounts without pooling data, a calendar notetaker bot, Start instant meeting, and Mia. Mia is the recruiter copilot, in-app and over a secure MCP server so recruiting can be run from Claude.
The candidate side
Mock interviews targeted at a company and a round, coding tests, resume feedback rewritten for one opportunity, prompt engineering practice, Psychometric Test, VC Pitch and a question bank. Plus the campus layer: placement ERP, cohort readiness, drives, offers and exports.
The platform edges
An OAuth-protected MCP server that puts the whole recruiting surface inside ChatGPT and Claude. White-label domains with automated TLS and DKIM-signed sending. A Chrome extension that masks PII, and a desktop app for strict proctoring.
Want the full picture first?
The about page covers where the company came from, how the platform fits together, and what is still in progress.
Read about Expert HireHow we work
Four habits, and one of them is being honest about what is not built.
Here is what the day looks like. You can check most of it against the product and this site before you talk to us.
Say the ship state out loud.
Look at how this site is written. Greenhouse is available today in early access. Bullhorn, Lever, Workday and Vincere are on the roadmap. SOC 2 and ISO are in progress. SSO and SAML are not here yet. We talk the same way internally.
Evidence beats opinion.
If you want to change a scoring behavior, bring the transcripts. If you want to change a flow, bring what candidates actually did in it. Strong opinions are welcome once they are attached to something.
Write it down.
The work runs across time zones, so the default medium is written. A decision that only exists in a call did not happen. Clear writing is most of the job, and it is a large part of what we read for when you send us something.
Ship for four audiences at once.
A change to the rubric touches an employer report, a staffing client workspace, a campus dashboard and a student practice score. You will not work on one audience in isolation. You do see the whole loop, from a student practicing for a company to that company running the round.
Where the work happens.
Expert Hire Inc. is registered in Dover, Delaware, United States, and the product was incubated at IIT Madras. The people we build for sit across the US, LATAM, the Philippines and India, so hours overlap rather than match. Interviews on the platform run in multiple languages for the same reason.
How hiring runs here
You go through our own product.
Every candidate here runs the same rounds our customers run, on the same rubric, producing the same reportcard. It is also how we find out where the product still hurts.
Your resume goes through a screening round.
The same resume screening round our customers run. It reads skills, experience and education against the role and produces a fit score. No interview involved. It is a first pass, not a gate, and a person still reads the borderline ones.
Resume Screening roundYou take an AI interview with Ethan.
A live voice round. Questions adapt to what you say. It runs 24/7 and in your language, so you book it around your current job instead of taking a morning off.
AI Interview roundYou do real work, not a puzzle.
Engineering roles get a coding test: a sandboxed problem, real execution against hidden test cases, scored on correctness, efficiency and code quality. Roles where AI fluency is the job get an AI prompt assessment, scored on 8 dimensions including prompt quality and iteration strategy.
Coding Test or AI Prompt AssessmentYou meet the team.
A human-led interview, hosted here or on Zoom, Meet or Teams with a notetaker in the room. Recorded and transcribed with your knowledge. It scores against the same rubric as every other round, so it sits alongside them instead of overwriting them.
Human-Led Interview roundA person decides.
Your rounds roll into one report: a score out of 100, a verdict line, a skill breakdown drawn from the transcript and the code. Someone here reads it and makes the call. There is no automated pass or fail at any point, and we do not auto-reject anyone.
Human review, alwaysWant to see the rubric before you sit any of it? The methodology page explains what each round scores and how the evidence is attached.
What you end up with
The same report a hiring manager gets.
Every round in our process produces this, and it is what our customers buy. Tabs across the top for Overview, Analysis, Submissions and Resume. Share and Download PDF in the header. Reports are never metered.
A score you can argue with.
The overall score sits next to a fit chip and a one-line verdict. Under it, the skill breakdown comes from the transcript and the code. If you disagree with a number, there is something specific to point at.
The evidence travels with it.
The recording plays inside the report, and the transcript is speaker separated and searchable with All, Candidate and Interviewer filters. Integrity status sits in the same view. Nothing about your round is hidden behind a summary.
Coded product UI. The candidate, the role and every figure in this screen are sample data.
See a sample report in fullWhat we look for
Craft, ownership and judgment, in that order.
People use our software to make decisions about someone's career. Here is what tends to work here and what tends not to.
What tends to work here
- You have shipped something end to end and can walk through the parts that went wrong.
- You write clearly enough that a decision survives without you in the room.
- You have an opinion about fairness in hiring and can hold it under pressure from a deadline.
- You are comfortable owning a surface rather than a ticket queue.
- You will read a transcript before arguing about a score.
What usually does not
- A list of frameworks with nothing built behind it.
- Wanting the AI to make the hard call so nobody has to.
- Needing a fully specified brief before anything starts moving.
- Treating candidate experience as somebody else's problem.
None of this is a filter on where you studied or how many years you have. The rounds are the same for everyone, and the report is what we read.
Where help is useful
Areas of work, not a list of openings.
We do not post roles we have not committed to. These are the parts of the platform where an extra pair of hands changes what we ship next. If one of them describes what you do, tell us which and why.
Evaluation and scoring systems
Skills, weightage and the job description feed the score. This is the work of making an AI interview, a coding test and a human-led round comparable, and keeping every score tied to the transcript and the code behind it.
Product engineering
Live voice interviews, sandboxed code execution, the ATS pipeline and multi-round workflows, the MCP server and white-label delivery.
Candidate experience
The practice side: mock interviews, resume feedback, question bank, Psychometric Test and VC Pitch. Also the part employers never see, which is what a rejection feels like.
Trust, integrity and compliance
Proctoring, malpractice signals that stay advisory, audit trails, PII masking, and jurisdiction work across NYC Local Law 144, Illinois, California and the EU AI Act.
Hiring and campus operations
Recruiter workflows, placements and productivity reporting, and the university side: placement ERP, bulk student onboarding, drives, offers and cohort readiness for a placement cell.
Go to market
Working with companies, staffing firms and campuses, running the product on their real roles, and bringing back the parts that did not fit.
Curious what the customer side of this looks like before you write? Companies, staffing firms, universities and job seekers each get their own page.The API and SDK live here.
How to apply
One email. No portal, no account.
Send a note. If the fit is there we start the rounds described above. If it is not, we say so instead of leaving you refreshing an inbox.
What to put in it
- Which area, and why that oneName the surface from this page you want to own. A specific answer tells us more than a cover letter does.
- One thing you shipped end to endWhat it was, what you decided, what broke, and what you would do differently. Links beat descriptions.
- How to reach youA resume or a profile link, and a time zone. Interviews run around your schedule, so we need to know what yours is.
Applying is not a commitment to a full process. Plenty of good conversations here have started with someone telling us what we got wrong.
FAQ
Working here, answered.
The questions people send us before they send a resume.
Still unsure whether to write? Ask first. Nobody here minds the question.
Get in touch
Tell us what you would own.
No roles posted does not mean no hiring. The conversation starts with you, and the process after it is the one we ask our customers to trust.