# Senior Product Manager \| Integrations & AI Foundations (f/m/d) at Bunch

> Germany · Senior · Product · open. Source: https://alderhq.app/jobs/ashby%3Abunch%3Ae8f7920c-332c-4363-a1ef-dedb8351b461 (read 2026-11-20).

## Facts (observed)

- Company: [Bunch](https://alderhq.app/companies/bunch)
- Location text: Berlin
- Where it can be done from (inferred): Germany (evidence: "Location: Berlin")
- Posted: 2026-11-04 (16 days ago)
- First seen by alder: 2026-11-04 · last verified: 2026-11-10
- Reposted: 0 times · edited: 0 times
- Apply on the employer's site: https://jobs.ashbyhq.com/bunch/e8f7920c-332c-4363-a1ef-dedb8351b461
- Skills listed (inferred): LLMs

## Pay and sponsorship

- Posted pay: not posted
- Market band for similar posted roles (Berlin, n 7): p25 €86k, median €121k, p75 €121k

## Description (the employer's text)

bunch is building the backbone of private markets. We are enabling next-gen fund operations with one integrated system that combines secure data infrastructure, AI-powered workflows and expert fund services. If you value ownership, growth through real responsibility, and working with a thoughtful, ambitious team, this role might be for you.

### Your Role

As Senior Product Manager for Integrations & AI Foundations, you're the product owner for bunch's integration and AI platform. You have two kinds of customers: external developers and fund-manager teams connecting to bunch, and the internal product teams building on top of what this team provides.

Integrations & AI Foundations is a new platform squad. It owns the infrastructure that lets data move securely between bunch and other systems, and lets every other product team build AI capabilities — extraction, retrieval, agents and evaluations — without each one solving the same hard problems again.

The team covers five connected areas:

Integrations / API layer — API contracts, authentication, versioning and deprecation, rate limits, uptime and SLA, and a Developer Portal

MCP layer — agent-callable tools built as thin wrappers over the same API contracts, including a customer-facing MCP offering

Document Parsing — a reusable ingestion, decomposition and extraction pipeline, with human-in-the-loop review and accuracy monitoring

Retrieval / RAG — chunking, embedding, vector store and retrieval, with access control enforced on every query, offered as a callable service

AI Agentic Layer — the agent framework, memory, guardrails, evaluation harness, observability and agent registry other teams use to ship their own agents

Document extraction, agent workflows, evaluations and multi-provider model fallback under EU data residency are already in production.

Fund managers don't want to retype data that already lives in their own systems, and they want to pull their bunch data into their own tools and their own AI. Customers and prospects ask for APIs, CRM and accounting syncs, automated exports of capital accounts, and MCP access.

You'll turn those signals into a coherent platform: one that customers integrate with, and that our own product teams build on, instead of a string of one-off builds.

### You will

Set platform strategy and sequencing across the five areas: decide how to slice the work and what ships first

Run APIs as a product: contracts, versioning and deprecation policy, rate limits, SLAs, Developer Portal and documentation, and time-to-integrate as a measure you actively drive down

Define and launch the customer-facing MCP offering: which capabilities get exposed, tool descriptions that agents actually invoke correctly, and a security review gate for anything customers touch

Generalise document parsing: the first use case is already live; your goal will be to take it over and generalise it so more teams can reuse its logic

Shape retrieval and the agentic layer as services other teams consume, with access control, guardrails, evaluations and observability built in rather than bolted on

Draw and hold ownership lines with domain teams. When a request doesn't map to a named responsibility, run the scoping conversation rather than absorbing it

Own platform-wide policy with Engineering and Security: data classification and residency, LLM cost and rate governance, and an integration and extraction registry so nothing gets built twice

Work directly with customers — fund managers, CFOs and their technical teams — and with Customer Success to understand integration needs, then turn them into platform decisions

Define success through goals and key results and clear metrics per area: API uptime and SLA, time-to-integrate, MCP adoption, extraction accuracy, and reuse across teams.

### What success looks like

### In your first 90 days

Resolve and document the MCP audience, governance model, and hand-off point between platform and domain teams

Agree a plan and timeline for moving extraction from the existing team onto the shared pipeline

---
Data: alder (https://alderhq.app). Cite https://alderhq.app/jobs/ashby%3Abunch%3Ae8f7920c-332c-4363-a1ef-dedb8351b461 and the date you read it. Observed facts come from the employer's own listing; location reach, level, family and skills are inferences.
