Guaranteed hired in 7–14 days —
or you pay nothing.
From a complete 35-page PRD to a live production MVP. Supabase + Lovable + Claude API — built as a demonstration of AI-first product management methodology applied to Romania's broken recruitment market.
Problem, solution, architecture, business model — the full product build from PRD to production.
A recruitment system
built to frustrate.
Romania's job market operates on bilateral frustration: candidates don't know why they're ignored, employers lose 10k EUR/month per unfilled position. No existing platform solves both sides end-to-end.
"My experience before building this: eJobs, BestJobs, LinkedIn. 120 applications. 3 responses. 0 offers. It wasn't a CV problem — it was a system problem." — insight from user research
75% of CVs never reach a human. Eliminated by algorithms candidates can't understand or navigate — before anyone reads their experience.
"BFS" — the Romanian term for applications that vanish without reply. Chronic anxiety, active job search abandonment, loss of self-confidence.
Manual screening takes weeks. HR is blocked in triage, not selection. Cost of opportunity is massive — key roles stay unfilled for 45–60 days.
Nine out of ten employers lose 54,000 RON per eligible hire simply because they don't know the government subsidy exists. Structural market failure.
165,000+ active candidates aged 40–64 in Romania. Recruitment market worth 500M EUR/year. Average time-to-hire: 45–60 days. Every day a role is unfilled costs the employer ~500 EUR. AngajatLaTermen reduces this to 7–14 days — guaranteed in contract.
Hired in 11 days.
Or you don't pay.
AngajatLaTermen solves both sides simultaneously: ATS bypass + micro-course for candidates, pre-verified candidates + ANOFM integration for employers. Success-based pricing — zero financial risk for both parties until contract is signed.
CV upload → automatic parsing → gap detection against available jobs. Zero upfront cost.
Not a 6-month bootcamp. A targeted course on the missing skill. Mandatory 80%+ quiz score to earn the badge.
CV sent directly to the hiring manager via email. Completely bypasses the ATS filter. No algorithm between you and the decision maker.
15 best-fit jobs selected by AI with compatibility scores. Employers receive only badge-verified candidates — no screening needed on their side.
Real project, paid. Employer sees actual work before committing. Candidate proves value. Both sides eliminate the risk of a bad hire.
Candidate: free — 0 RON. Employer: monthly subscription for job posting + interview scheduling with pre-verified candidates, plus 10% success fee on annual salary (minus 54k RON ANOFM if 45+).
Success-based pricing eliminates the primary sales objection — financial risk. The candidate doesn't pay if they're not hired. The employer doesn't pay if they don't hire. This structure reduces onboarding friction to zero and enables a 10-second pitch: "Hired in 7–14 days. Or 0 RON."
Serious no-code.
Enterprise-grade.
A demonstration that an AI Product Manager can ship a complete, secure, production-ready MVP without a dedicated development team. Supabase for enterprise backend, Lovable for rapid UI, Claude API for the intelligence layer.
Or 0 RON.
No ATS black holes. No ignored CVs. No "we'll be in touch." Verified micro-course, direct access to the hiring manager, paid trial. You pay only when you sign the contract.
12-table schema, Row-Level Security per role, JWT authentication, AES-256 encrypted CV storage, auto-generated REST API. 34 SQL queries written and tested in production.
Candidate dashboard, employer panel, real-time KPI admin panel — built through iterative prompt engineering. Full AngajatLaTermen design system implemented.
Skill gap analysis from CV, job matching with compatibility scoring (15 personalized recommendations), fraud detection, CV parsing to structured JSON.
Row-Level Security in Supabase, per-role access policies, Storage Policies for CV uploads, audit log for CV access, GDPR + EU AI Act compliance by design.
Idempotency keys for duplicate payments, escrow logic for success fee, multi-gateway fallback (Netopia/PayU), 3× retry logic on failure with exponential backoff.
54k RON subsidy eligibility check, retry logic on timeout, manual admin override, ATS bypass via direct hiring manager email (AWS SES).
12 PostgreSQL tables: polymorphic users (8 roles, 11 states), candidate_profiles, employer_profiles, jobs, courses, enrollments, interviews, paid_trials, contracts, payments, badges, applications. Full state machine for both flows (candidate + employer) with explicit transitions and edge case handling for race conditions, duplicate payments, ANOFM timeouts, and ghost candidates.
Unit economics
impossible to ignore.
Cost per placement: 127 RON. Revenue per placement: 22,000 RON. Margin: 99.4%. Success-based pricing eliminates risk for both sides and creates a model that closes itself in the sales conversation.
Total cost: 127 RON — micro-course production 50 RON + email API 2 RON + trial payment processing 15 RON + admin support 60 RON. Revenue per placement: 22,000 RON — 0 RON candidate (free) + 10% employer on average gross salary 6,500 RON × 12 months. Employer also pays a monthly subscription for job posting access and interview scheduling with pre-verified candidates. Each additional placement costs near zero marginal cost. Margin: 99.4%.
Candidates pay nothing — zero upfront, zero success fee, zero cost at any stage. The platform is entirely free for job seekers. Revenue comes exclusively from the employer side.
10% of annual gross salary, paid only when the employment contract is signed. If candidate is 45+, ANOFM subsidy of 54,000 RON is automatically applied — effectively making the hire free or net positive. 95% margin.
Employers pay a monthly subscription to post job listings and schedule interviews with pre-verified, badge-holding candidates. Access to the verified talent pool, interview calendar, and candidate management dashboard. Tiered by volume: Starter / Professional / Enterprise.
MRR target Month 6: 200k RON. Net profit margin: 62%. Revenue streams: Employer Subscription (monthly, tiered by volume — Starter / Professional / Enterprise), Success Fee Employer 10% (95% margin), Trial Commission 15%, ANOFM Processing Fee 500 RON/candidate. Candidates use the platform entirely for free.
AI PM in action.
Skills proven through artifacts, not declared on a CV. From user research to production deploy.
vs. 45–60 day Romania market avg
REST spec · 5 groups · Production
With real frequency data from research
127 RON cost · 22,000 RON revenue
Projections based on PRD v1.0 · Validation data: Perplexity AI user research · Romanian labor market 2024–2026
Data-driven validation of every product assumption — before the first line of code was written.
Real data.
Clear decisions.
Primary research conducted with Perplexity AI on Romania's labor market — 9 validated pain points, 4 B2B segments mapped, 3 actionable personas, and a complete Jobs-to-be-Done framework. All produced before writing a single line of product spec.
Structured queries on public sources: HR studies, ANOFM reports, eJobs/BestJobs statistics, INS Romania data, industry publications. Verified against multiple independent sources.
Focus on segments 25–64 years actively on the job market, and SMEs + enterprise with active hiring needs in Romania. Both sides of the marketplace.
9 problems validated with real frequency data, 4 B2B segments with distinct needs, direct input for PRD v1.0 pricing model and feature prioritization.
Research identified a new B2B segment (Recruitment Agencies), exposed a critical SME segmentation gap, and re-attributed ghosting as an employer problem — all before MVP build started.
9 problems.
Real frequencies.
Every problem documented with occurrence frequency and primary impact. The data fed directly into the Problem & Opportunity section of PRD v1.0 and drove feature prioritization for the MVP sprint.
Research vs. assumptions.
What changed.
Every hypothesis from PRD v1.0 put face-to-face with research output. Green = confirmed. Amber = gap identified. Purple = new insight not covered in the original PRD.
Recruitment Agencies — a segment entirely absent from PRD v1.0 — have the highest volume (100+ hires/year) and the exact need the platform solves: pre-verified candidates + ATS bypass + API access. This is both a direct revenue channel and a distribution partnership opportunity.
4 segments.
4 distinct needs.
Research revealed that the B2B market is not homogeneous. Each segment has a different hiring frequency, a different core problem, and a different optimal package. Treating them identically means zero conversion across the board.
1–3 hires / 12 months
Small budget — can't afford a traditional headhunter (20–25%)
Guaranteed, fast, cheap. Zero upfront.
Urgent hire, no internal HR function
5–15 hires / year
Bulk hiring + strict SLA + HR team with limited bandwidth
7–14 day guarantee, 3–7 curated candidates per role
Rapid team growth, scaling operations
50+ hires / year
API integration + white-label + complex reporting for board
5-day SLA, 10+ trials/month, integration with own ATS
Cost-per-hire exceeding budget + board pressure
100+ hires / year per agency
Done with manual screening — human resources are expensive
Pre-verified candidates + ATS bypass + API access
White-label + revenue share — becomes a distribution channel
3 personas.
Built from data.
Personas derived directly from validated pain point frequencies. These aren't invented archetypes — they are composite profiles based on the segments with the highest occurrence rates in research.
78% CV rejection by ATS + 48% age discrimination. His experience is an asset the system treats as a liability. ANOFM subsidy makes him attractive to employers rather than a risk.
65% of candidates blocked by the upskilling barrier. Wants a fast career transition, not a long program. The platform is completely free for her — zero cost at any stage, including after getting hired.
68% of employers lose 10k EUR/month per unfilled role. Unaware of ANOFM (90% of employers). Double win: speed guarantee + automatic 54k RON subsidy he didn't know existed.
What they actually
need to accomplish.
JTBD framework: "When [situation], I want to [motivation], so that [outcome]." Each job mapped to the exact platform feature that resolves it. No feature without a job. No job without a validated frequency.
When I send CVs and receive no response, I want to know my CV reaches a real human, so that I stop losing months in a void with no feedback.
ATS Bypass — direct email to hiring manager + video CV. "Job-Ready Verified" badge as a quality signal that bypasses automated filters entirely.
When I'm rejected as "overqualified," I want access to employers who see my experience as an asset, so that I'm hired because of it, not in spite of it.
ANOFM integration — employers receive 54,000 RON subsidy for 45+ hires. Mihai's experience becomes a financial advantage for the employer, not a risk signal.
When I want to change fields but lack the required skills, I want to acquire them quickly and without upfront cost, so that I can apply next week, not in 6 months.
Micro-course 3–5 days, 3h/day, 0 RON upfront. The platform is entirely free for candidates — no fee at any stage, before or after hiring.
When I apply for jobs, I want to know the salary before investing time in 3 interview rounds, so that I don't waste 2 weeks on a process that leads to an unacceptable offer.
Mandatory salary transparency in job listings — salary_min / salary_max fields visible before applying. No hidden compensation until round 3.
When a role has been vacant for 2 months and I'm losing money daily, I want qualified candidates in maximum 7 days, so that I stop paying the cost of an empty seat.
7-day SLA guaranteed in contract. 3–5 pre-verified candidates with badge. 2–3 day paid trial eliminates the risk of a wrong hire before full commitment.
When I hire people over 45, I want to access available government subsidies, so that the real cost of the hire drops significantly below standard agency fees.
Monthly employer subscription gives access to the verified candidate pool + interview scheduling. 10% success fee only on contract signed. For a 45+ hire at 6,500 RON salary: employer fee is 10% = 7,800 RON minus 54,000 RON subsidy = net positive hire.
What changes
in PRD v1.1.
Research confirmed 6 hypotheses and opened 3 concrete adjustments. These will be incorporated into PRD v1.1 before the next development sprint.
1,200+ agencies in Romania with 100+ hires/year each. Need: pre-verified candidates + API + white-label. Model: Enterprise + revenue share. This is not just a client segment — it's a distribution channel that can 10× the platform's reach without proportional marketing spend.
PRD v1.0 treated SME as one segment. Research shows that 10–50 employee companies (cheap + fast) and 50–200 employee companies (SLA + bulk) have fundamentally different needs and willingness to pay. Separate pricing tiers, feature sets, and sales motions required.
70% of employers suffer from candidates disappearing after interviews. PRD v1.0 listed ghosting under candidate pain points. Correct reallocation: this is an employer reliability problem. Solution: escrow commitment feature for verified candidates + auto follow-up sequences on the employer dashboard.
90% frequency of ignorance means every B2B employer is a qualified lead for ANOFM integration. This needs to move higher in the pitch and demo flow — it's the differentiator most immediately monetizable. "We'll automatically recover 54,000 RON you're currently leaving on the table" closes faster than any other value prop.
65% of candidates are blocked by the upskilling barrier. The 4-day / 0 RON upfront micro-course is the direct answer to the primary objection. Message "4-day course. Not 6 months." is confirmed by research as the winning framing. This positioning should lead all B2C acquisition copy.
A PRD without user research is documented fiction. This data changed 3 product decisions before the first line of code — and saved weeks of building the wrong thing.
Want to discuss a similar project?
Product strategy, AI integration, or no-code MVP — let's book a 30-minute call.