AI API for recruitment in India
Resume screening, job descriptions, and candidate matching powered by unoblox's ₹-native AI API — one key, every model, billed monthly in rupees.
Recruitment in India is plagued by volume and bias. Thousands of resumes for one role; screening by hand is slow and subjective. AI can triage resumes, generate objective job descriptions, and match candidates to roles—but only with transparent, auditable decision logic. unoblox offers models with explainable reasoning, all billed in rupees on your invoice.
Resume screening & ranking
Qwen3 Max (₹120.95 / ₹604.77) reads resumes and ranks candidates by fit against a job description. Analyze skills, experience, education, and red flags (employment gaps, frequent job-hopping) in one pass. Recruiters review the top 10; AI narrows the funnel from 1000 to 10 in minutes. Use for consistent, bias-conscious screening.
Job description generation
Claude Sonnet (₹201.6 / ₹1008) writes compelling job descriptions from a role brief: title, team, responsibilities, must-haves, nice-to-haves. It auto-adjusts language tone for different seniority levels (entry-level vs. director). Regenerate for multiple talent pools without rewriting from scratch.
Candidate skill matching
DeepSeek V3.2 (₹26.21 / ₹38.30) analyzes both resume and job description to score skills overlap. "Candidate has Python, Django, PostgreSQL; role needs Python, FastAPI, MongoDB. Match: 70% (core Python yes, frameworks different)." Help recruiters prioritize candidates with rare but critical skills.
| Use-case | Recommended model | ₹ input/output | Why |
|---|---|---|---|
| Resume screening | Qwen3 Max | ₹120.95 / ₹604.77 | Comprehensive skill analysis |
| Job descriptions | Claude Sonnet | ₹201.6 / ₹1008 | Clear, compelling writing |
| Skills matching | DeepSeek V3.2 | ₹26.21 / ₹38.30 | Precise overlap detection |
| Embeddings: similarity | Embeddings | ₹0 (freemium) | Candidate clustering |
Frequently asked questions
Will AI screening introduce bias? AI can reduce bias vs. human screening if trained on unbiased data. But all AI reflects its training data. Audit: score diverse candidates equally? Are salary ranges gender-neutral? Use AI as a funnel, not a final decision.
Can I reject candidates based solely on AI screening? Not legally. Use AI to rank; humans review the top N and make offer/reject decisions. Document your process for labor compliance.
Which model understands Indian education and certifications best? Claude Opus (₹504 / ₹2520) has broad knowledge of IIT, BITS, ICAI certifications. Qwen3 Max is solid and cheaper; test both on your local hiring market.
How do I avoid AI screening disadvantaging non-traditional backgrounds? Adjust scoring: give credit for bootcamps, freelance projects, open-source contributions, not just formal degrees. Curate few-shot examples that value diverse paths.
Can I use embeddings to find similar past hires? Yes. Embed resumes of your top performers; when a new candidate arrives, find similar profiles. This personalizes screening toward your high performers.
Is there a free tier to test resume screening? Yes. Qwen3-1.7B (₹0) does basic resume parsing. Upgrade to Qwen3 Max for ranking and matching complexity.
Get started in rupees → https://unoblox.ai/sign-in
Start building in rupees
Call every major model through one OpenAI-compatible endpoint, billed in ₹ on a GST invoice.