Choose timezone
Your profile timezone:
Q-SOLVE Hackathon 2026 - Kenya Edition
→
Africa/Nairobi
Strathmore University
Strathmore University
, , , , , , , , , , ,
Description
Pre-Hackathon Sessions:
The sessions from 24–28 August 2026 are held online. Participants should attend the sessions relevant to their assigned track. Sessions marked Mandatory are required for all participants.
Main Hackathon:
The in-person hackathon will take place from 1–3 September 2026 at Strathmore University, Nairobi, Kenya.
-
-
Introduction to the Agriculture mentors, review of the expanded challenge briefs, and methodology brainstorming.
Required attendees: Agriculture Track participants only
Google Meet: https://meet.google.com/prf-gwkz-mvv
Calendar: https://calendar.app.google/aZ8ugpK7QEsTqcHT8Speakers: Billy Nyaga (Space Fika), Dr Dickson Owuor (Strathmore University), Dr Getrude Alworah (KALRO-Coffee Research Institute), Mr Kisilu Wambua (Penda Kujua)Meeting Minutes: QSOLVE Agriculture Track (SDG 2) Mentor Session
Date: August 24, 2026
Meeting Type: Mentor and Participant Kickoff
Organizers: QSOLVE Hackathon Organizing Committee
1. Attendance
Mentors & Organizers:
-
Billy Nyaga (Organizer)
-
Dr. Getrude Alworah (Mentor - Agricultural Data)
-
Dr. Dickson Odhiambo (Mentor - Computer Science/Facilitation)
-
Mr. Kisilu Wambua (Mentor - Quantum Computing)
Participants Present:
-
Adika Awino
-
Avery Inyangala
-
Bruce Wamburu
-
Evans Nzomo
-
Kennedy Mutugi
-
Mercy Amondi
-
Peter Kimani
-
Samuel Muoria
-
Samuel Waweru
-
Tracy Ratia
-
Ruth Ngotho
-
Enoch Mureti
-
Stacy Wangui
-
Akram Ali Omar
Participants Absent (With Apologies - WA):
-
Silvia Mganga
2. Meeting Objectives
-
Connect mentors with participants.
-
Clarify the objectives and technical requirements for the Agriculture Track.
-
Define the specific hackathon challenges and data sources.
-
Finalize team distribution strategy.
3. Discussion Summary
3.1. Hackathon Track Definitions
The Agriculture track comprises two primary challenges. Mentors will facilitate based on their expertise (Computer Science, Agriculture, Quantum Computing).
-
Track 1: Fertilizer Distribution Optimization (CMDVRP)
-
Goal: Optimize supply chain networks to minimize transportation costs and ensure timely fertilizer delivery.
-
Constraints: Modeled as a Capacitated Multi-Depot Vehicle Routing Problem. Constraints include vehicle capacity (10 tons/200 bags), a 10-day dispatch period, and an 8-hour drive limit per route.
-
Data Sources: Open Street Map (via OS-Mx Python library) for road networks; Kilimo.com (1981-2025) for yield metrics to estimate demand.
-
Note: The provided equation for transport costs can be modified with sound justification consistent across both classical and quantum models.
-
Track 2: Maize Yield Forecasting
-
Goal: Develop a system to accurately predict maize production to assist government planning. This is a yield forecasting problem, not weather forecasting.
-
Constraints: Predictions must simulate a forecast date of June 30th. Strict rules against data leakage—no data post-dating June 30th (e.g., harvest records) can be used.
-
Data Sources: CHIRPS rainfall dataset (converted from TIFF/BIL to CSV/shapefiles); NDVI (via Digital Earth Africa/Jupyter notebooks) as a proxy for temperature. Additional maize-producing counties should be included to expand the training set.
3.2. Technical Modeling & Quantum Requirements
A core principle for all challenges is that participants are not required to prove quantum superiority. The goal is rigorous comparison.
-
Hybrid Approach: Implement a real agricultural problem, develop a credible classical baseline, implement a quantum-classical hybrid solution, and fairly compare them.
-
Track 1 Quantum Execution: Use classical methods for network clustering, then solve sub-problems using quantum algorithms (CUBO/Ising models). Scale up from smaller instances.
-
Track 2 Data Splitting:
-
Training: Data up to 2018 (All preprocessing/feature selection must use training data only).
-
Validation: 2019–2021 data.
-
Testing: 2022 onwards data.
-
Track 2 Quantum Execution: Limit inputs to 4-12 meaningful features. Execution on a Cube simulator is mandatory (hardware is optional).
-
Evaluation Metrics: Mean Absolute Error, Root Mean Square Error, and R-squared improvement over the classical baseline.
3.3. Team Distribution Restructuring
To accommodate the 15 participants across 3 teams and ensure all teams perform both classical and quantum analysis, the challenge distribution was revised. Track 2 (Maize Forecasting) will be split into two distinct sub-projects using different datasets.
The finalized topics for the three groups are as follows:
-
Group 1: Fertilizer Supply Chain Optimization (Track 1)
-
Objective: Formulate a Capacitated Multi-Depot Vehicle Routing Problem (CMDVRP) to optimize the fertilizer distribution network, minimizing travel distance and transportation costs while meeting demand and adhering to constraints (e.g., 10-ton capacity, 8-hour limit).
-
Group 2: Maize Yield Forecasting – Dataset A (Track 2, Sub-project 1)
-
Objective: Develop a predictive model to forecast maize production (in tons per hectare) using CHIRPS rainfall data and vegetation indices up to a strict cutoff date of June 30th, applied to a specific set of maize-producing counties.
-
Group 3: Maize Yield Forecasting – Dataset B (Track 2, Sub-project 2)
-
Objective: Tackle the same maize yield prediction challenge as Group 2, using the same June 30th cutoff and variables, but applied to a different assigned dataset or region to ensure a separate working environment.
(Note for all groups: Every team must create an analysis-ready dataset, implement a classical baseline, build a quantum-hybrid alternative, and compare performance across at least 5 test instances.)
4. Action Items & Next Steps
-
Attend Qbraid Training Session: All participants must join the session on August 25, 2026, at 7:00 PM for quantum platform training.
-
Manage Team Distribution: Billy Nyaga will finalize participant groupings, circulate them to mentors, and establish communication channels immediately.
-
Document Rules: Kisilu Wambua will consolidate all technical constraints, "dos and don'ts," and modeling requirements into a single guide to share with organizers.
-
Construct Analysis-Ready Data: Hackathon teams will perform ongoing data gathering (OS-Mx, Kilimo, CHIRPS, Digital Earth) ensuring models have necessary inputs without data leakage.
-
Implement & Compare Models: Hackathon teams will formulate classical baselines and quantum hybrid solutions, testing them across increasing instances and comparing them using defined metrics throughout the hackathon.
-
Project Submission: Hackathon teams will submit comprehensive documentation at the end of the hackathon, including code, provenance logs, Cube notebooks, error analysis, limitations, and a non-technical PowerPoint.
-
-