AI-native Product Manager

Short description

QR Datalab creates an AI agent for institutional investors capable of autonomously analyzing public companies and generating investment ideas. We are looking for an AI-native Product Manager who not only manages the product, but also prototypes the features himself using Claude Code, Cursor and agentic workflows. You will be a bridge between investment expertise and AI technologies, turning complex analytical processes into working agent systems.

Responsibilities

• Creating a product vision and roadmap for an AI agent: what investment tasks do we automate, how do we prioritize, how do we measure success • Rapid prototyping of new features using Claude Code, Cursor, and Windsor — from an idea to a working prototype in days, not weeks • Designing agentic workflows: which agents are needed (analyst agent, data agent, backtester agent), how they interact, where is the person-in-the-loop • Working with LLM API directly: writing and optimizing prompta, experimenting with reasoning chains, testing various models (Claude, GPT-4, o1) • Conducting user research with hedge fund analysts and portfolio managers: understanding their workflow, pain, and analysis quality requirements • Creating product specs and user stories for the dev team, but with a deep understanding of the technical implementation • Building the metrics framework: how to measure the quality of an agent's investment ideas, accuracy predictions, user engagement • A/B testing of various prompt strategies and agent architectures • Coordination between investment analysts (domain experts) and engineers — translation from the "language of finance" to the "language of AI"

Requirements

• 10+ years of experience in product management, preferably in fintech, B2B SaaS or AI products • Hands-on development experience with Claude Code, Cursor, Windsor or similar AI coding tools • Ability to write Python code yourself: FastAPI endpoints, working with LLM APIs (Anthropic, OpenAI), basic data analysis • Deep understanding of LLM agentic workflows: tool calling, multi-agent orchestration, ReAct pattern, chain-of-thought, memory systems • Knowledge of frameworks: LangChain/LangGraph, LlamaIndex, CrewAI or the experience of building custom agent systems • The ability to quickly create working prototypes: from the idea of a feature to a demo in 1-2 days • Understanding the basics of investment analysis (or rapid learning) — what is DCF, comp analysis, how to read 10-K • Experience working with technical stakeholders and the ability to speak the same language as ML engineers • Strong analytical skills: the ability to formulate hypotheses, design experiments, interpret results • English — free reading of tech/product documentation • Self-driven: the ability to work in conditions of uncertainty 0→1 product

It will be a huge plus:

• Experience launching AI products with LLM from idea to production • Background in finance or work experience in investment tech • Knowledge of modern RAG architectures and vector databases • Experience with prompt engineering frameworks (DSPy, LMQL) • Understanding MLOps: how to deploy and monitor LLM systems • Portfolio of AI projects on GitHub

Conditions

• Competitive salary + equity (share in the company) — you will become a co-builder of the product • Remote work or hybrid from Astana Hub with flexible schedule • Direct work with founder, flat structure, quick solutions without bureaucracy • Unlimited budget based on AI API (Claude, GPT-4, o1) and premium tools (Cursor Pro, GitHub Copilot) • Access to financial data (Bloomberg, FactSet) to understand domain • Freedom to experiment: try new models, frameworks, approaches • The ability to create a category-defining product in intersection AI × Finance • Participation in AI/fintech conferences and networking with the hedge fund community

Contact information

Email: akhmet.assyl@gmail.com

When responding, attach: • A link to GitHub/a portfolio with examples of AI projects or prototypes • An example of a prompt/agent workflow that you created • A case study: how would you design an agent to analyze quarterly earnings calls?