About the role
VectorShift is building the institutional memory layer for private markets. We work with private equity and private capital firms to turn the data scattered across their spreadsheets, fund admins, and documents into a living system: portfolio monitoring, fund analytics, and AI agents that answer the questions deal teams and CFOs actually ask. Our design partners include leading mid-market PE firms, and we ship weekly.
The role
You'll be the engineer who sits closest to our customers — embedded with PE fund finance teams and deal teams, building with them, and carrying what you learn back into the product. The job is threefold:
1. Build agents and workflows for clients and pilots. Configure and build AI agents, data pipelines, and automations on the VectorShift platform for customers — for both sides of the firm. For the finance team: ledger ingestion and reconciliation, quarterly data collection, document extraction (fund admin statements, LPAs, valuation backups). For the deal team: agents that prep the Monday portfolio review, extract underwriting assumptions from IC memos and board decks, rank exits by impact on fund returns, and answer "what happens to the fund if we sell this company next year?" on their real ledger. You own the pilot from first data file to the moment partners are asking your agents questions in a live meeting.
2. Bake it into the product. Every engagement teaches us what private-markets customers need out of the box. You'll turn one-off builds into reusable product: seeded templates, standard ingestion pipelines for common document formats, default configurations for fund structures (co-invests, SPVs, blockers, bridge loans). Your north star is making the next deployment need less of you.
3. Build the one-off features customers need. When a client needs something the product doesn't do yet — an export format, a calculation variant, an integration with their fund admin — you scope it, build it, and ship it, working directly with our core engineering team. You'll have real commit access to the product, not a sandbox.
