Software Engineering Intern

Job description

About

Abstrakt is building the AI-native workspace for media intelligence .

We started with a precision monitoring engine: an AI system that reads thousands of articles a day so comms and PR teams do not have to. You give it a topic and an angle; it returns exactly what matters, by meaning, not keywords. Selected by AI, validated by human experts.

Now we are going further: from monitoring to intelligence. A workspace where communications teams ask questions and get sourced answers (did the press pick up our key messages? who owns the narrative?), shaped together with design partners.

· 
Based at Station F , Paris (largest startup campus in Europe).

· 
Small, founding team (Julien, CEO; Marvin, CTO).

· 
Partners: UNESCO, Havas, Microsoft for Startups, Station F, Wilco.

· 
We serve PR / comms agencies and enterprises, in France and beyond.

Early-stage startup: you see the whole product, ship to production fast, and your work has direct, visible impact.

Job Description

The role

You work directly with Marvin (CTO) on Abstrakt's technical platform: the pipeline that discovers, extracts, qualifies and synthesizes the news, and the products that deliver it (the human-validation back office, the mobile app, the workspace).

Over 6 months, you could:

· 
Ship features end to end (backend-heavy, some frontend), from design to production.

· 
Work on our AI workflows: relevance, sentiment and narrative qualification, summarization, synthesis. We run a hybrid engine: LLM APIs for deep reasoning, in-house small language models (SLMs) for speed and the things LLMs are bad at.

· 
Push automation forward: part of the pipeline still needs a human touch today; help close that gap (data cleaning, deduplication, article selection).

· 
Improve data quality and coverage across sources and languages (100+ countries).

· 
Help build the workspace and the mobile app.

· 
Strengthen the foundations: tests, observability, internal tooling.

You will not be boxed into one corner. The point is the full picture, from crawl to client.

Our stack

· 
Backend: Python, Django, Django REST Framework.

· 
Async & queues: Celery, RabbitMQ.

· 
Data: PostgreSQL, vector search.

· 
AI / ML: in-house SLMs, LLM APIs, clustering, classification, embeddings.

· 
Frontend: Alpine.js, Tailwind.

· 
Infra: Docker, Docker Compose, Nginx.

You do not need to know all of it. We want someone strong in Python, curious about the rest, and able to learn fast.

What you get

· 
Real ownership: your code ships to production and serves real clients.

· 
Direct mentorship from the CTO and broad exposure: backend, AI, infra, product.

· 
A seat at Station F , in the heart of the Paris startup ecosystem.

· 
A real problem worth solving: applying AI to something concrete and useful (access to quality information), not a demo.

Practicalities

· 
Duration: 6 months. Start: as soon as possible.

· 
Location: On-site at Station F, Paris.

· 
Compensation: about 670/month (French internship legal minimum).

· 
Perks: Navigo pass (transit) and Swile lunch card.

· 
Internship agreement (convention de stage) required.

Preferred Experience

· 
CS / engineering student looking for a 6-month internship (convention de stage required).

· 
Strong in Python ; experience with Django (or another web framework) is a real plus.

· 
Genuine interest in AI / NLP and the drive to get hands-on.

· 
Rigor and autonomy: here, consistency beats cleverness. We write clean, readable code built to last, not for a weekend.

· 
You ship, and you take ownership of what you ship. Use every tool at your disposal, AI included, but understand the systems you build and stand behind your work. We value a genuine drive to understand how things work, not only to make them work.

· 
Intellectual honesty: you flag what is wrong, you ask questions, you do not make things up.

· 
French is not required. We speak a lot of French at the office day to day, but our code, docs and technical work are in English, and you can do the whole job in English: we are glad to switch anytime. French is a nice plus (the product is French-first), never a barrier.

Nice to have (not required): data engineering, information retrieval, NLP, SQL / query optimization, a bit of DevOps.

Recruitment Process

We keep it short and move fast: you hear from us quickly at every step.

· 
Intro call (~30 min, video) with Marvin (CTO): your background, what you want out of the internship, and a first technical chat. Mostly mutual fit.

· 
On-site at Station F (about 2-3 hours): a live pair-programming session on a small, realistic problem (we care how you think, debug and ship, not whether you memorized algorithms), plus a chat with the team and a look at where you would work.

We aim to get back to you with a decision within one to two weeks of the first call.

Additional Information

·  Contract Type: Internship (Between 6 and 7 months)
·  Location: Paris
·  Salary: 670 / month