Senior Data Engineer – Live Ingestion & Reliability

Note: This job is not for Fomogo, but one of our clients, NextAlphaAI About NextAlpha We build AI-powered intelligence for India's retail investment market. Our product, InvestorAI, is embedded inside broker apps — giving investors company intelligence and portfolio alerts drawn from BSE/NSE filings. Seed-funded, post-stealth, integrating with broker partners now. We're building a governed AI platform — not an agent playground. Every number the AI shows an investor traces back to a verified filing. If you've built pipelines that failed, recovered, and stayed correct, this role is for you. What you'll work on • Ingest BSE/NSE financial filings at scale — quarterly results, annual reports, shareholding patterns, corporate actions. Structured CSV and unstructured PDF • Extract structured data from PDFs and investor presentations — including documents where numbers appear only in tables or charts. This is the hardest part of the role • Enforce a validation and approval pipeline before any data reaches the AI or investors. Nothing bypasses the human review gate • RAW → APPROVED → QUARANTINED lifecycle, full lineage, immutable audit log. No silent failures, no partial updates leaking downstream • Corrections and restatements via versioning, not overwrites • Own the embedding pipeline for semantic search — chunk, embed, store, query • Dead letter queues, idempotent retries, checksum verification — day one, not afterthoughts What we're looking for • 3–5 years on production data ingestion pipelines — not analytics, not ML modelling • Real PDF extraction experience in production: pdfplumber, pymupdf, or equivalent • Strong Python, solid SQL, schema design • Pipeline orchestration — Airflow, Prefect, or equivalent • Correctness-first instinct. You reason about failure modes before shipping Nice to have: Financial data exposure (BSE/NSE formats, fintech/wealthtech background). Vector store experience (pgvector, Pinecone). Multi-tenant environments. This role is NOT a BI, streaming, delete-and-reload, or ML role. It is a data reliability and ingestion ownership role. The pipeline you build controls what the AI can say — and what it cannot. Why join Small team, real product, real broker partners. High ownership, senior guidance. Your work directly controls what investors see — and what the AI is not allowed to say.