What You’ll Learn
- Create tools that provide actionable insights for human decision-making
- Master the development cycle from CLI tools to interactive dashboards
- Work with industry-relevant datasets and technologies
- Build portfolio-worthy projects demonstrating AI-augmented systems
Choose Your Focus
Explore different angles based on your interests and expertise:
AI vs Traditional • UX Design • Model Optimization • Real-World Integration
| Project & Domain | Sprint 1 (CLI) | Sprint 2 (Streamlit Dashboard) |
|---|---|---|
| FinNews Pro Finance |
Sentiment-Price Correlation Analyzer with statistical significance output | Live trading signals, AI-generated market rationale, and confidence intervals |
| SecuChat 2.0 Cybersecurity |
Alert Prioritizer with ATT&CK mapping | ATT&CK technique visualizer, investigation steps, and shift handoff reports |
| AutoSense Edge Manufacturing |
Predictive Maintenance Estimator (failure timing + confidence) | Degradation timeline, repair vs. replace optimizer, and engineer feedback logs |
| GeneGPT Lite Genomics |
Variant Classifier with ACMG scoring | Clinical report drafts, filterable tables, and evidence visualizations |
| MoodBeats AI-Human Interaction |
Biosignal-Based Mood Classifier | Real-time mood tracking, AI-curated playlists, and mood evolution feedback |
| StudyAI EdTech |
Flashcards & Quiz Generator with API performance comparison | Adaptive study assistant, spaced repetition scheduler, and API benchmarking dashboard |
AI Augmentation Types:
■ Rule-Based ■ Probabilistic ■ Knowledge-Enhanced
Deliverable Requirements:
- Sprint 1 — Jupyter Notebook or equivalent
- Sprint 2 — Streamlit or equivalent dashboard
Next Steps:
Step 1: Overview
Step 2: You Are Here
- Browse the 6 AI-augmented hackathon projects.
- Identify the projects that match your skills and interests.
- Continue to Step 3 to start your prep modules and get qualified.
Up next: Step 3 – Prep Modules & Qualification
1. FinNews Pro: AI-Driven Market Narratives
Domain: Finance
Key Resources:
- NewsAPI – News Headlines JSON API
- yfinance – Yahoo Finance Python Wrapper (Unofficial but widely used)
- Prophet – Open-source Time Series Forecasting Library
Project Sprints:
Build a sentiment-driven financial analysis system:
Sprint 1: CLI tool that flags statistically significant correlations between news sentiment and stock price movements.
$ python finnews.py --ticker AAPL --days 7
Output: "Negative sentiment spike correlated with AAPL price drop (p=0.04)"
Sprint 2: Streamlit dashboard with
- Live trading signals + confidence intervals , with forward-looking price trend forecasts.
- AI-generated “Market Pulse”: Key sentiment drivers and alternative scenarios
GenAI Integration Examples:
- Sentiment analysis of news articles
- Market narrative generation from multiple sources
- Risk scenario creation based on current trends
Bonus Challenge:
- Implement false-positive tracker for failed signals
- Add causal chain visualization (News Event → Price Move Lag → Sector Impact)
2. SecuChat 2.0: SOC Triage Automation
Domain: Cybersecurity
Key Resources:
- CIC-IDS2017 Intrusion Detection Dataset
- MITRE ATT&CK Framework (for mapping attack techniques)
Project Sprints:
Build a security operations center assistant:
Sprint 1: CLI tool prioritizing alerts by severity and mapping them to MITRE ATT&CK techniques.
$ python secuchat.py --alerts-file alerts.csv
Output: "Top 3 Alerts: SQL Injection (Critical), DDoS (High), Brute Force (Medium)"
Sprint 2: Streamlit dashboard showing
- ATT&CK technique mapping with AI-suggested investigation steps
- Shift handoff reports with executive summaries
GenAI Integration Examples:
- Alert enrichment and contextualization
- Investigation playbook generation
- Executive summary creation from technical alerts
Bonus Challenge:
- AI-suggested initial containment actions (e.g., auto-isolate affected endpoints)
- False-positive feedback loop (“This was benign” → model retraining)
3. AutoSense Edge: Predictive Maintenance for Smart Factories
Domain: Manufacturing
Key Resources:
Project Sprints:
Build a predictive maintenance system:
Sprint 1: CLI tool predicting failure timing from sensor patterns with probabilistic confidence.
$ python autosense.py --sensor-data turbine14.csv
Output: "Engine 14: Predicted failure in approximately 8 days (high confidence)"
Sprint 2: Streamlit dashboard with sample features:
- Interactive degradation timeline
- “Repair now ($5K) vs. replace later ($25K)” scenarios
- Engineer override logs (“Disagreed with AI on 07/15 - faulty sensor”)
- Model confidence indicators
GenAI Integration Examples:
- Anomaly explanation for non-technical staff
- Maintenance recommendation narratives
- Automated reporting for management
Bonus Challenge:
- Optimize model size for resource-constrained edge devices
- Add ROC curve evaluation on held-out samples to verify model robustness.
4. GeneGPT Lite: Multi-Gene Variant Interpretation Assistant
Domain: Genomics
Key Resources:
- ClinVar VCF Files (GRCh38)
Optional resources for deeper dives: - gnomAD – Allele Frequency Reference Database
- COSMIC – Somatic Mutation Reference (Lite Download)
- Ensembl Variant Effect Predictor (VEP)
Project Sprints:
Build a genomic variant interpretation system:
Sprint 1: CLI tool for variant annotation and ACMG pathogenicity scoring
$ python genegpt.py --variant BRCA1:c.68_69delAG --acmg
Output: "Pathogenic (PVS1, PM2, PP3)"
Sprint 2: Streamlit dashboard with
- Filterable variant classification tables
- AI-generated clinical report drafts dynamically assembled from retrieved evidence (RAG-powered)
- Evidence visualization (population frequency, clinical significance)
GenAI Integration Examples:
- Clinical report generation from variant data
- Literature summarization for novel variants
- Patient communication templates
Bonus Challenge:
- Benchmark the classification system against a held-out ClinVar test set
- Literature retrieval + summarization for novel variants (integrate into RAG pipeline)
5. MoodBeats: AI-Powered Music from Biosignals
Domain: AI-Driven Human Interaction
Key Resources:
Project Sprints:
Build an emotion-driven music recommendation system:
Sprint 1: CLI tool classifying emotional states from biosignal data
$ python moodbeats.py --input eeg_sample.csv
Output: "Detected Mood: Relaxed → Recommended Playlist: Ambient Chill"
Sprint 2: Streamlit dashboard example: Mockup by Claude
GenAI Integration Examples:
- Mood state narratives from EEG patterns
- Music selection explanations
- Weekly mood trend summaries
Bonus Challenge:
- Add personal calibration option (e.g., first session fine-tunes model to the user’s baseline signals).
- Implement “emotional journey” planning for workout or study sessions
6. StudyAI: AI-Powered Study Tools Generator
Domain: EdTech
Key Resources:
- OpenAI API – GPT for Summarization and Question Generation
- Anthropic Claude API – Safe and Cost-Efficient LLMs
- Cohere Generate & Summarize APIs
Project Sprints:
Build an adaptive study assistant:
Sprint 1: CLI tool generating flashcards and quizzes with API benchmarking (accuracy, cost, behavior comparisons).
$ python studyai.py --input biology_chapter.pdf --output flashcards
Output: "Generated 20 flashcards + 10 quiz questions on Photosynthesis"
Sprint 2: Streamlit dashboard example: Mockup by Claude
GenAI Integration Examples:
- Content generation across multiple APIs
- Performance benchmarking and reporting
- Adaptive content difficulty adjustment
Bonus Challenge:
- Implement cross-API scoring reports to evaluate performance consistency across providers.
- Add visual concept mapping for interconnected topics