VenturePulse AI: Deep-Tech Sector Early-Movement Detection and Forecasting

Project

Identify deep-tech sectors that are about to move based on early public signals, before conventional databases capture them.

Objectives and deliverables

• Build a market-shift engine using arXiv, GitHub, patent and funding feeds.
• Develop an explainable and tunable momentum score for each sector.
• Perform a historical backtest to assess whether the signals predicted past breakouts.
• Create dashboards and briefings with heatmaps, sector radar views and cited reports.

Key challenge

Avoid lookahead bias by enforcing strict point-in-time data collection from the first week.

Target

Beat a funding-volume baseline on precision and recall in the backtest.

Technical stack

Python, PostgreSQL, D3.js, PyTorch, Pandas, LangGraph, Scikit-learn, FastAPI and Next.js.

Candidate profile

AI engineering or data science.

Internship conditions

• 1 intern.
• Full-time, on-site internship in Monastir, Tunisia.
• Available for long-term work after the internship.
• Willing to learn and take on the challenge.
• Relevant skills and interests.
• Adequate written and spoken English.
• Ability to work side by side with people from diverse backgrounds and disciplines.

Application documents

Apply with an English CV and a cover letter. The application deadline is 25 November.

Advantages

• Monthly bonus according to profile.
• Achievement bonus.
• In-demand training.