The language model that runs the interview.
Tulifo is an AI research and product company. We train our own models to run fair, verifiable interviews for any role, in any language.
Welcome. I’m Xiya, and I’ll be interviewing you today.
- Identity matched
- Live camera
- One face in frame
- Editor input
- Window focus
- 00:00Candidate joinsXiya introduces the role and explains how the session works.
- 00:02Identity checkedFace match, live camera and one person in frame, confirmed before the first question.
- 00:16The interview beginsStructured questions from your JD, asked and answered in the candidate’s language.
- LiveCoding, if neededA shared editor for any programming language, with input and focus signals recorded.
- AfterReport and BGVScores against your rubric, with background verification in the same file.
Latest releases
View allBackground verification, inside every interview
BGV now runs alongside the Xiya interview and lands in the same report.
Read moreIntroducing the verified candidate pool
Candidates interview and clear background checks once, then get matched to employers.
Read morepinmbo is live
A learning companion for kids, built on Chittu, that gives hints instead of answers.
Read moreAt Tulifo, we build AI so every candidate gets a fair interview and every employer can trust the result.
A hiring decision rests on a conversation. Today that conversation is scheduled across weeks, run by tired panels, and mostly forgotten once it ends.
We think the interview should be the most reliable signal in hiring. That takes a model trained for the job: one that asks follow-ups, hears many languages and scores against a rubric people can read.
So on Tulifo the AI interview is the whole interview, not a first-round screen. People review the result and make the final decision.
One model. Six ways to meet it.
Two products serve employers, three serve people looking for work, and two are about how the model itself learns.
A fully autonomous video interview for any role, in the candidate’s language, with background verification built in.
See openings that match the skills you already have, and apply in one step.
Candidates who have interviewed and cleared background checks in advance, shown to employers whose JD fits.
Rehearse a Google-style or Microsoft-style loop with interviewers from top tech companies.
A learning companion that teaches step by step and gives hints instead of answers.
Our own language model, with open weights published on Hugging Face.
Research
Our researchInterview-specific training
Genuine fine-tunes for interviewing are rare. We train for follow-up questions, rubric scoring and fair pacing, not general chat.
Beyond Western English
Vernacular and non-Western-English speech is where most interview models fail. It is where we focus.
Integrity signal fusion
Identity, camera, presence, editor and focus signals combined into one picture of a session’s integrity.
Public reliability benchmarks
We believe interview models should be measured in the open, and we are working toward publishing how ours perform.
Built for hiring across borders.
Candidates interview in the language they think in, in any time zone, for any role. Prices are flat and in US dollars, wherever you hire.
Language coverage depends on the role. Tell us where and in which languages you hire, and we will confirm.
Trust is part of the product.
SecurityConsent first
Candidate consent is recorded and versioned. Nobody is shown to an employer without agreeing to it.
Integrity on record
Identity, camera, faces in frame, editor input and window focus are logged with every interview.
People decide
Xiya recommends and explains. A person on your team makes every hiring decision.
Certification path
We are working toward ISO 27001 first, then SOC 2 Type II.
Put Xiya on your next hire.
Share a job description and invite your first candidates today. Your team reviews every result and makes every decision.