At Raven, we are building AI assistants for manufacturing plants which can offer intelligent assistance for day-to-day operations—like training, job planning, troubleshooting, reporting and also helps with real-time analytics. We’re taking decades of manufacturing expertise and combining it with AI to solve real operational problems. We’re a small, focused team working to make industrial operations safer, smarter, and more efficient.

We’re based in Bangalore, building for teams that work in some of the toughest environments.

About Us

Raven is a YC-backed startup (S22) building AI assistants for manufacturing plants. We’re taking decades of manufacturing expertise and combining it with AI to solve real operational problems. Backed by top VCs, we’re a small, focused team working to make industrial operations safer, smarter, and more efficient. We’re based in Bangalore, building for teams that work in some of the toughest environments.

What You’ll Work On

This internship is about pushing the boundaries of applied AI for complex industrial data. You’ll:

• Experiment with open-source and commercial AI models across vision, OCR, and LLMs.
• Fine-tune models on custom datasets and evaluate their performance on real-world tasks.
• Prototype workflows that combine perception (symbols, text, layouts) with reasoning (graphs, classifications).
• Help design and run benchmarks to measure accuracy, efficiency, and deployment readiness.
• Work closely with the core team to translate research findings into production-ready components.

Who We’re Looking For

• Strong Python skills and experience with PyTorch or TensorFlow.
• Prior exposure to training or fine-tuning ML models (academic projects or internships).
• Curiosity and persistence to try multiple approaches, analyze results, and iterate fast.
• Comfortable working in ambiguity, where problems are open-ended and datasets are messy.
• Bonus: familiarity with YOLO/Ultralytics, MMDetection, Detectron2, Hugging Face, or OCR libraries.

Duration

• 3–6 months internship, with potential for extension or conversion to a full-time role.

Why Join Us

• Work on high-impact AI problems in a space that’s still largely untouched by modern ML.
• Access to real industrial datasets (documents, diagrams, logs) not easily available elsewhere.
• Learn how to take ideas from research notebooks → deployed systems in weeks.
• Collaborate closely with a small, technical founding team in a YC-backed, VC-funded startup.

• Design and improve AI infrastructure that scales across industrial use cases.
• Build retrieval-augmented generation (RAG) systems to make sense of complex operational data.
• Create and optimize multi-modal pipelines to process information from documents, diagrams (e.g., P&IDs), and live data streams.
• Develop evaluation systems to measure model performance in real-world industrial contexts.
• Work on knowledge graphs and other tools to link operational and maintenance data across systems.
• Build systems for real-time data processing from equipment sensors and logs.