Pehchaan
Product Design Consultant · AI hiring platform
Pehchaan is an AI hiring platform. Candidates build a profile, take skill tests and do an AI interview. Recruiters get a report on each candidate. I designed both sides. The big decision: show recruiters the evidence first, and the AI score last, so a person, not a number, makes the hiring call.
Context
Background
Hiring tools have improved a lot. Applicant tracking is faster, interviews moved online, and AI now screens resumes. But recruiters still decide on very little: a resume, a few notes and a gut feeling.
Pehchaan wanted to change that. Instead of one resume check, it collects evidence over time: the resume, skill tests, an AI interview and a growth plan. Recruiters see all of it in one place.
My role
My role
I designed the product end to end, working with the founder, product, AI engineers and frontend and backend engineers.
- Candidate app: onboarding, resume upload, profile, skill tests, growth plan and the AI interview.
- Recruiter app: dashboard, candidate list, candidate reports and the hiring pipeline.
- AI reports: how the AI explains what it found, with transcripts and answers to each question.
- Design system: shared components and patterns across both apps.
The problem
Three problems we set out to solve
Candidates can’t show what they can do
A resume lists experience. It doesn’t show how someone thinks, talks or learns.
Recruiters piece it together
Resume, tests, interview notes and emails all sit in different places. Nothing joins them up.
Nobody trusts a bare AI score
Most tools show a number. Very few say why. Recruiters don’t act on what they can’t explain.
So the question became: how do we help recruiters understand a candidate before asking them to judge one?
Candidate app
A home that says what to do next
The first home screen looked like every job site: applications, notifications, recommended jobs, profile completion, messages. In testing, people kept asking the same thing: “what should I do next?” Every card was shouting at once.
We replaced it with one path, in order: Resume → Skill tests → Growth plan → AI interview → Jobs. Jobs come last, once the candidate is ready, not first.
Candidate app
The resume fills in the profile
Most platforms store the resume as a PDF. Recruiters then read it by hand. Here, the candidate uploads their resume once and AI fills in the profile: education, experience, skills, projects. The candidate types less and can edit everything. The same data then feeds the tests and the interview.
Candidate app
Skill tests and a growth plan
People usually leave a job app once they’ve applied, and come back only for the next job. So instead of “apply to more jobs”, Pehchaan asks: what can you get better at before your next opportunity? The skill tests show where someone stands. The growth plan shows what to work on. Each step they finish adds evidence recruiters can see.
Candidate app
An AI interview that doesn’t feel like an exam
“AI interview” makes people nervous. Our first version had a countdown timer, a big record button and long instructions. It felt like an exam.
So before the first question, the intro answers everything people worry about: how many questions, how long, can I practise, what happens next, how am I judged. Interviews are also split into short rounds, so a company can pick the rounds a role needs instead of one long session.
Recruiter app
A dashboard built around decisions
A recruiter handles dozens of jobs and hundreds of applicants. The first dashboard was full of charts: total applications, funnels, team activity. It looked impressive, but recruiters come to hire, not to read reports.
We put what needs action first: jobs that need attention, then candidates waiting for review, then interview progress. Reports are still there, just lower down. Every block on the page has to help with one decision, or it goes.
The candidate list works the same way. Each row shows who the person is, how far they’ve got, how they scored and how urgent they are, so recruiters don’t have to open every profile to decide who to interview.
Recruiter app
Evidence first, score last
An early report opened with “Overall score: 86%. Strong hire.” Everyone who saw it asked the same thing: why 86? The AI was giving an answer without a reason, and recruiters were left just watching.
We flipped the order: summary, strengths, areas to improve, each question with its answer, the transcript, and only then the recommendation. By the time recruiters reach the score, they already know why.
AI and people
AI helps, people decide
Hiring needs context, judgement and a sense of fit. So we drew a clear line between the two.
AI does
Reads the answers, spots patterns in how people communicate, summarises interviews, points out strengths and concerns, and puts the evidence in order.
Recruiters do
Read the context, judge team fit, run follow-up interviews, weigh the trade-offs and make the final call.
Trade-offs
What we chose not to build
- Onboarding: not every field up front. Candidates give the minimum and add more as they go.
- Candidate home: not a dashboard of everything. One path from resume to jobs.
- AI: not candidate → score → hire. AI organises the evidence, people decide.
- Recruiter dashboard: not more charts, even when stakeholders asked for them. Only what needs action.
- Scope: no networking, communities, social feeds or badges. Every feature had to make a hiring decision clearer.
Outcome
What changed
For candidates
A shorter start, a clear next step, and a reason to come back and improve, not just apply.
For recruiters
One place for all the evidence on a candidate, and AI reports they can check and trust.
For the product
One system linking candidates, AI and recruiters, with shared patterns ready for new features.
Reflection
Reflection
I thought I was designing a hiring platform. I was really designing a way to make decisions with confidence. A good interface isn’t one that looks modern. It’s one that helps people decide well, and that applies to any product that asks people to trust it.