Real scoped projects from Bangalore AI startups. Build your portfolio, earn ₹2K–25K per task, and get hired by the companies you work with — while keeping your main job.
Complete agent-guided learning sessions. Earn verified skill badges through a 5-minute conversational assessment — not a quiz. Verified badges unlock bounty eligibility.
Agent-verifiedFilter bounties by skill, duration, and payout. Each listing shows the exact tech stack required. If your verified badges match — apply in one click. No cover letter needed.
Skill-matchedScoped, time-boxed deliverables — 1 to 30 days. No open-ended commitments. Submit your work through Scorio. Company verifies, Scorio releases payment. 20% commission on completion.
Protected paymentCompanies hire the people they have already worked with. Every completed bounty is a live portfolio entry. After 3 successful projects with the same company, direct hire conversations begin.
Direct hire pathwayBuild a production-ready FastAPI service wrapping our Llama fine-tune. Needs batched inference, request queuing, and Prometheus metrics.
Replace a 40-field form with a conversational Claude agent that extracts KYC data through natural dialogue and validates in real time.
Implement RAGAS evaluation on our 12K support ticket dataset. Measure faithfulness, context precision, and answer relevance across 3 chunking strategies.
Implement scaled dot-product attention and multi-head attention as a clean educational notebook with annotations and test suite.
Build a weekly Airflow DAG that scrapes 10K job postings, runs NER for skills/role/salary, deduplicates, and loads into BigQuery.
Use Claude API to generate Playwright test assertions from UI screenshots. Cover 15 critical user journeys. Integrate with GitHub Actions.
Scorio curates good-first-issues from AI projects that Bangalore companies actually use. Close a promotion gap, build your portfolio, earn a ₹500 runner-up reward.
Indian language translation models. Contribute to Kannada, Tamil, Telugu, and Hindi NLP benchmarks. High-impact for Indian AI infrastructure.
The most widely used LLM framework. Good-first-issues for documentation, test coverage, and Indian language model integrations.
LLM evaluation framework. Contribute new metrics, improve documentation, or add support for Claude and Gemini models.
Data framework for LLM applications. Issues range from new data connectors to retrieval improvements and agent memory systems.
Nobody stays at Micro-tasks forever. Each completed project unlocks higher-value work, better company access, and eventually a direct hire conversation.
Per completed bounty. No subscription required to earn. No upfront fees. You only pay when you get paid.
Every applicant has agent-verified skills — not just a Udemy certificate. Companies get deliverable-ready professionals, not aspirational candidates.
Free to join. No subscription needed. 47 bounties open right now.