
QUANTUM AI HACKATHON - AI agents meet qubits, powered by XpyQ
βοΈπ€ Quantum AI Hackathon β Where AI Agents Meet Qubits
Build Next-Generation Quantitative Trading Systems across HFT, Mid-Frequency Trading, Risk & AI Foundation Models
Presented by Berkeley Gateway Accelerator, XpyQ, Oakland Hardware Meetup, and Informal Spaces β this is an intensive, in-person weekend challenge bringing together quant researchers, quantum algorithm engineers, ML practitioners, and AI model builders to prototype the next generation of trading systems.
π Berkeley Gateway HQ β Downtown Berkeley
π¨ How to Secure Your Spot as a Participant
Two steps are both required to be considered:
1οΈβ£ Apply on Luma to reserve your seat β Register here
2οΈβ£ Complete the follow-up survey sent after registration β this allows organizers to match you to the right track and team
Spots are limited to approximately 10 teams. Apply early.
π Observers are also welcome!
π Why This Matters Now
The next era of quantitative finance is not classical. U.S. quant funds suffered their worst momentum selloff in three years in early 2026 β dropping 2.8% in just two weeks. Meanwhile, the quantum-AI risk hedge platform market is expanding at a 28.5% CAGR, projected to reach $10.71B by 2030. The window to build domain expertise before institutional standardization narrows closes within the next 2β3 years.
McKinsey projects that 50% of major financial institutions will deploy hybrid quantum trading systems by 2028. This hackathon is a live preview of the teams building that future.
π Four Frontier Competition Tracks
β‘ Track 01 β High-Frequency Execution
Deploy Quantum Reinforcement Learning (QRL) agents on live Limit Order Book (LOB) data. Your benchmark: beat the LSTM/Transformer baseline.
π Track 02 β Mid-Frequency Alpha Generation
Leverage QAOA/VQE portfolio optimization across 50β500 assets. Your benchmark: beat Markowitz + CVXPY.
π‘οΈ Track 03 β Risk & Regulatory Compliance
Build hybrid CVaR circuits combined with quantum Monte Carlo methods. Your benchmark: beat classical historical VaR.
π§ Track 04 β AI Foundation Models (Cross-cutting)
Fine-tune Chronos, TimesFM, or Kronos, or deploy an agentic LLM pipeline. Your benchmark: beat zero-shot generic TSFMs.
π€ Who Should Attend?
π Quant Researchers & Algorithmic Traders
Classical factor crowding is structurally fragile β quantum-enhanced strategies tackle regime failure from angles classical models simply cannot reach. Build systems judged on real metrics: Sharpe ratio, CVaR, information coefficient, and execution P&L.
π¬ Quantum Engineers & ML Practitioners
Work against real financial data β not toy benchmarks β with full hardware access to IBM Quantum (Heron chip), IonQ, AWS Braket, and D-Wave, plus frameworks including PyTorch, JAX, LangChain, and LlamaIndex. Results are evaluated by industry professionals who prioritize reproducible outcomes.
πΌ Investors & Hedge Fund Guests
Join as an observer, judge, or sponsor. A formal Awards Dinner the following Wednesday provides a structured opportunity for industry guests to engage with top-performing teams.
π Event Agenda
ποΈ Prep Week β 7 Days Prior (Async via Slack)
| Day | Topic |
|---|---|
| Monday | Market microstructure primer: LOB dynamics, tick data, order types |
| Tuesday | Quantum computing primer: QAOA, VQE, QUBO, NISQ-era constraints |
| Wednesday | AI foundation models for finance: Chronos, TimesFM, Kronos β fine-tuning vs. zero-shot |
| Thursday | Quantum ML: QRL, QSVM, VQC, quantum kernel methods |
| Friday | Dataset access, API keys (IBM Quantum, Alpaca, Polygon.io), track briefings |
Arrive Day 1 with your environment configured and classical baselines already running.
π¨ Day 1 β 12-Hour Build Sprint
- 9:00 AM β Doors Open
- 10:00 AM β Opening Keynote: Quantum + AI for Hedge Funds β The 2026 Landscape
- 11:30 AM β Build Block 1: Core algorithm prototyping, quantum circuit design, model architecture
- 1:30 PM β Lunch + Mentor Office Hours
- 2:00 PM β Build Block 2: Backtesting, benchmarking, hyperparameter tuning
- 5:30 PM β Checkpoint Demos (5 min/team, non-judged) + Mentor Feedback
- 6:00 PM β Build Block 3: Integration, result validation, final documentation
- 9:00 PM β π Code Freeze β all repositories locked
π€ Day 2 β Presentations
- 9:00 AM β Doors Open, final submission upload & presentation setup
- 10:30 AM β Team Presentations: 10-min demo + 5-min judge Q&A per team
- 12:00 PM β Judge Deliberation (closed session)
- 12:30 PM β Participants dismissed; scores sealed until the Awards Dinner
π Awards & Prizes
| Award | Prize |
|---|---|
| π₯ Grand Prize | $2,000 + exclusive invite to Formal Awards Dinner |
| π₯ Runner-Up | Second overall + exclusive invite to Formal Awards Dinner |
| π Track Champions Γ4 | Best project per track: HFT, MFT, Risk & Compliance, AI Foundation Models |
| βοΈ Quantum Advantage Award | Most rigorous demonstration of quantum speedup or accuracy lift over classical baseline |
| π€ Best Foundation Model | Most impactful FinTSFM or agentic LLM deployment |
| π‘ Most Creative Approach | Judges' discretionary prize for novel problem framing |
π οΈ Infrastructure & Tools
βοΈ Quantum Hardware & SDKs
IBM Quantum (Heron chip) Β· IonQ Β· AWS Braket Β· D-Wave
Qiskit Β· PennyLane Β· Amazon Braket SDK Β· D-Wave Ocean
π€ AI & Foundation Models
Chronos Β· TimesFM Β· Lag-Llama Β· Kronos Β· MOIRAI-2
LangChain Β· LlamaIndex Β· Hugging Face Transformers
π Classical ML & Data
PyTorch Β· TensorFlow Β· scikit-learn Β· JAX
Polygon.io Β· Alpaca Markets Β· yFinance Β· Simulated LOB tick feeds
β‘ Execution Simulation
Backtrader Β· Zipline-reloaded Β· Custom HFT sandbox
π€ Partner Organizations
- Berkeley Gateway Accelerator β Venture accelerator focused on frontier capital and deep tech commercialization, based in Downtown Berkeley
- XpyQ β Building next-generation quantum-AI infrastructure for quantitative finance applications
- Oakland Hardware Meetup β A Bay Area community of hardware engineers, makers, and builders advancing hands-on technical collaboration
- Informal Spaces β Flexible venue and community infrastructure for technical events, workshops, and collaborative builds
π Important Notes
- Teams may self-pair or be matched from the accepted-but-unmatched applicant pool
- Prep materials distributed one week prior via Slack
- An IP participation agreement is required at registration
π Register on Luma
Register to See Address
Get directionsScan with your camera β the event opens in the Somo app.









