MyStartupWave / PS 04
HyperSpeed Quick Commerce
Disruptive logistics and ordering systems for minutes-level delivery at scale
Improve quick-commerce operations through a reproducible simulation and measurable gains over a baseline.
HyperSpeed Quick Commerce
The objective
The challenge.
Quick-commerce operators, dark-store managers and riders must keep deliveries fast and economical as demand grows. At peak times and during rain, stock-outs or rider shortages, nearest-store routing and first-available rider assignment can create delays, cancellations and uneven workloads. Forecasting, picking, inventory and dispatch are often optimised separately, leaving little visibility across the complete order journey. Design a logistics or ordering innovation that improves the end-to-end operation, and prove its impact through a realistic, reproducible simulation. Around 10-minute delivery promises provide the industry context; the evaluation targets below define the required improvement.
Operations teams and dark-store managers, with delivery riders, customers, inventory planners and replenishment teams as stakeholders.
Fulfil more orders within their promised delivery window, especially during peaks and disruptions, while improving delivery time, rider utilisation and stock availability without increasing cost per order.
Product capabilities
What to build.
Required features
- One clear operational innovation
Innovate in at least one of five directions: demand forecasting and inventory placement, order batching, rider dispatch and routing, dark-store picking, or the customer ordering experience. Integrate it into an end-to-end order flow. You do not need to innovate in all five.
- A realistic city-zone simulator
Model at least 3 dark stores, 30+ riders and thousands of orders per day, including realistic road travel times, demand peaks and SKU-level inventory.
- A fair baseline
Implement a baseline policy, such as nearest store plus first-available rider, in the same simulator. Compare both systems on identical scenarios.
- Live operations visibility
Build a real-time dashboard showing orders, store load, rider positions and status, SLA breaches and key performance indicators.
- Adaptive decisions
Respond in real time to demand surges and disruptions such as rain, rider shortages and store stock-outs.
- Customer delivery estimates
Predict accurate delivery ETAs and show them to the customer when an order is placed.
Optional ideas
- Predictive positioning
Pre-position inventory or riders ahead of forecast demand.
- Stock-out recovery
Split orders across stores or suggest substitutions when items are unavailable.
- Rider-aware assignment
Include fairness and fatigue in rider allocation.
- Ordering prototype
Add smart recommendations and delivery-slot nudges in a customer-facing prototype.
- Real open maps
Use an actual open road network, such as OpenStreetMap data for an Indian city.
Typical workflow
Customers place orders and see ETAs → the system assigns, batches and routes orders → riders receive optimised tasks → operators monitor the dashboard and intervene when alerts fire.
Demo & deliverables
Show it working.
- Comparable simulation runs
Run the baseline and proposed system on the same simulated day, including at least one demand surge and one disruption.
- Live dashboard demonstration
Show the complete ordering, picking, dispatch and delivery flow, including the system’s adaptive response.
- Evidence of improvement
Present a side-by-side KPI comparison and explain why the improvement occurs. Provide reproducible scenarios that judges can rerun.
Success criteria
What success looks like.
Achieve at least 15% lower average delivery time OR at least 20% higher on-time, within-SLA delivery rate against the baseline on the same scenario, without increasing cost per order.
Make dispatch and routing decisions in under 2 seconds per order at 500+ concurrent active orders. Keep ETA mean absolute error below 2 minutes and make results reproducible.
Quantify cost per order and orders per rider-hour, and present a credible pilot plan. Show better rider utilisation, more reliable ETAs and a higher share of orders meeting the promised SLA.
Demonstrate end-to-end operation and adaptive responses without unsafe rider speeds or workloads. Respect workload and break limits throughout the simulation.
Scope & constraints
Build within the brief.
Your logistics or ordering innovation, simulator, baseline comparison, operations dashboard and ETA model.
Live-rider deployment, payments, procurement contracts and physical hardware. Drone or robot concepts may be discussed but must be simulated. A full customer app is optional, not required.
One city zone and a limited catalogue, such as 200–500 SKUs, are sufficient. You may approximate travel times from open road networks. You will need suitable map data, optimisation libraries and simulation compute.
Technical & safety requirements
Any technology stack is allowed. Run the decision engine in real time inside the simulation and use fixed random seeds so judges can repeat the scenarios.
Do not use live data from a quick-commerce company’s private systems. Location and customer records must be synthetic or anonymised.
Keep simulated speeds within realistic urban limits and respect road-safety and labour norms. Delivery gains must come from better planning.
The system should run on a standard laptop or free cloud tier. The dashboard must run in a browser.
Use open-source optimisation and mapping tools where possible. Document any paid API usage.
The impact estimate assumes simulated demand and traffic represent real operations. State those assumptions; a successful simulation does not itself establish real-world performance.
Resources & help
Useful before you start.
OpenStreetMap, OSRM or GraphHopper, Google OR-Tools, SimPy and other APIs available to your team are permitted. Tools and API credits are not provided by the partner; permitted services are used at your own discretion and cost.
Generate synthetic order, inventory and rider data, or use public grocery/delivery datasets and open road networks within their licences. Build your own simulation environment; no sandbox is provided.
Google OR-Tools vehicle-routing documentation, OpenStreetMap and OSRM documentation, and research on dynamic vehicle routing and last-mile delivery.
