Do you see patterns where others see noise? Can you turn messy data into AI that actually moves the needle? Do you get a kick out of shipping — not just analyzing? If yes, we'd love to work with you.
At Stylumia, we're looking for high-agency Data Science / AI Engineer interns to help build the intelligence layer behind how global fashion, beauty, and home brands make decisions.
🔍 What you'll do:
✨ Build and ship AI-powered features that decode trends across fashion, beauty & home
📊 Work with large-scale product, image, and consumer data to surface real insights
🧠 Prototype with LLMs, agents, computer vision, and recommendation systems
🛠️ Own problems end-to-end — hypothesis → data → prototype → production
📈 Run experiments, measure impact, and iterate fast
🤝 Partner with product, design, and business teams to turn models into outcomes
🔁 Continuously learn, break things, fix them, and level up
🧠 What we value:
Strong fundamentals in Python, SQL, and statistics
Hands-on experience building with modern AI — LLMs, agents, RAG, multimodal models
Comfort with the current AI tooling stack (OpenAI / Anthropic APIs, vector DBs, orchestration frameworks, evals)
Bias toward action: you'd rather ship a v1 than perfect a v3
High ownership — you don't wait to be told what to do next
Clear thinking, clearer communication
High learnability — the stack changes every quarter, and you enjoy that
Traditional ML framework experience (PyTorch, TensorFlow, scikit-learn) is a plus, not a requirement — we care more about what you can build today than what you've trained from scratch.
📩 To apply:
Send us a simple AI project you've built — could be a weekend hack, a tool you use yourself, a demo, anything. A working link or short demo video matters more than a polished resume. Tell us what you built, why, and what you'd do differently next time.
This role is for someone who wants to build real AI systems, see them used by the world's biggest brands, and grow at the pace the industry is actually moving.
Let's shape the future of fashion decision-making, together.
.jpeg)
