Founding Data QA and Evaluation Lead Offline

Altss is building the most advanced intelligence platform for alternative assets — extracting and enriching millions of investor, fund, and startup profiles from unstructured data. We already outperform legacy platforms like PitchBook and Fintrx in speed, visibility, and dataset depth.

Now we’re looking for a top 1% QA leader to own data quality, enrichment validation, and LLM output evaluation across our stack. You’ll work directly with the founder, infrastructure engineer, LLM engineer, and parser team — and design the full QA layer from scratch.

This is not a manual tester role. You will operate as the final checkpoint for intelligence quality, ensuring Altss becomes the most trusted LP intelligence product on the market.

What You’ll Own

• Lead QA for 100,000+ structured profiles across LPs, funds, startups, principals, and mandates
• Design validation rules for entity resolution, duplicate detection, and schema consistency
• Review and score low-confidence records (automatically flagged by LLM/infra)
• Evaluate LLM enrichment: bios, firm descriptions, investment strategies, summaries
• Build systems to detect hallucinations, mismatches, or missing entities
• Work closely with Infra + LLM engineers to close the QA → feedback → improvement loop
• Build QA dashboards, trackers, and scoring reports (manual + automated)
• Define quality standards and acceptance thresholds for Altss going forward

What We’re Looking For

• 5+ years in structured data QA, enrichment validation, or analyst-quality control
• Experience with high-volume, high-stakes data (e.g., investor databases, fintech, search platforms)
• Deep attention to detail and proven ability to spot inconsistencies and weak enrichment
• Familiarity with VC, PE, family office structures (check size, stage, principals, mandates, etc.)
• Hands-on experience reviewing or guiding LLM-generated outputs
• Bonus: experience at Crunchbase, CB Insights, PitchBook, Fintrx, or similar
• Bonus: familiarity with GPT-based evals, LangChain eval, or prompt QA chains
• Comfortable working with JSONL, Google Sheets, GPT, and internal QA tools

Why This Role Is Critical

• You’ll be the final reviewer of our intelligence output — the person who ensures clients trust what they see
• You’ll design the QA process, not inherit one
• You’ll have direct impact on the velocity and reliability of our product
• You’ll help us build a product that clients prefer over billion-dollar incumbents

Stack You’ll Interact With

• LangChain, OpenAI, Mistral (LLM enrichment layer)
• Prefect 2.0 (workflow + confidence routing)
• Postgres, DuckDB, JSONL (raw + structured data layers)
• GPT-4, Notion, Google Sheets (QA tooling)
• Internal dashboards, embedding-based entity linking