EN

EN

Internship – Software Engineering

Design and implementation of an evaluation framework for a financial simulation LLM agent (Fintech)

📍 Paris (hybrid) – Preferred start: between July and September 2025

Who we are

Apana is a next-generation fintech dedicated to wealth management professionals.
We are building a suite of products designed to simplify the daily work of advisors, family offices, and private banks:

  • A SaaS wealth management platform used by hundreds of professionals

  • An intelligent conversational agent that assists advisors with analysis, recommendations, and client interactions

  • White-label simulators designed to generate qualified leads

Our mission: to make wealth management decisions clearer, smarter, and more human by combining deep industry expertise with thoughtful technology.

Internship context and responsibilities:

Working directly with our CTO (Robin), you will operate at the core of our conversational agent’s intelligence to lay the foundation for its continuous improvement.
Your role: design and implement an automated evaluation framework (LLM evals) to objectively measure the agent’s performance and help it improve.

Your main tasks:

  • Define the structure of the evaluation dataset

  • Build an initial reference dataset (“golden dataset”)

  • Develop automated evaluation tools

  • Run evaluation campaigns

  • Identify weak points and contribute to improvement strategies

What we’re looking for

  • Master’s level student in Computer Science, Engineering or Artificial Intelligence (4th or 5th year)

  • Solid knowledge of Python

  • Interest in language models (LLMs) and their evaluation

  • Rigor, analytical mindset, critical thinking

  • Bonus skills: LangChain / LangGraph, automated testing, finance or interest in fintech

  • Motivation to contribute to a production-grade product at an early stage of AI structuring

Technical environment

Agent layer:
Python, LangChain / LangGraph frameworks for agent development
Likely use of LangSmith during the internship
Experimentation in Jupyter Notebook, test automation via Pytest

Backend:
Python API using Django + Django Ninja
PostgreSQL database

Frontend (not relevant to this internship):
React-router in framework mode

What we offer

You’ll join a growing startup with a compact and experienced team, led by founders who previously built a successful tech company.

The internship offers a high level of autonomy and responsibility, with the opportunity to work on real-world LLM challenges applied to a concrete use case, in a rapidly transforming industry (Fintech).

Practical information:

  • Duration: 4 to 6 months (flexible)

  • Preferred start date: July / September 2025

  • Location: Paris (hybrid)

📩 Want to join us?

Send us your resume, LinkedIn profile, and a quick note about what you’d like to learn from this experience at contact@apana.ai