Application Track

Build Something Real

For Developers, Engineers, and Anyone Who Ships

Theories are nice. Proofs are elegant.

But does it work?

This track is for people who build. Deploy a mission. Spawn ants. Make the colony solve a new problem.

The Challenge

Build a working mission on the Ants at Work platform.

Not a simulation. Not a prototype. A real mission running on real infrastructure:

• TypeDB Cloud (the colony’s brain)
• Agentverse (where ants run)
• Real pheromone dynamics
• Measurable results

Mission Ideas

Supply Chain Optimization

Model a supply chain as a graph. Products are resources. Routes are edges. Let ants find efficient paths.

Success metric: Reduce simulated logistics cost by 10%+

Code Navigation

Model a codebase as a graph. Files and functions are nodes. Dependencies are edges. Let developer navigation patterns become pheromone.

Success metric: Predict navigation 70%+ of the time

Research Paper Discovery

Model citations as a graph. Papers are nodes. Citations are edges. Let reader paths become trails.

Success metric: Surface relevant papers faster than baseline

Social Network Analysis

Model a social network. Users are nodes. Interactions are edges. Let engagement patterns reveal communities.

Success metric: Discover communities without labels

Game AI

Model a game state as a graph. States are nodes. Actions are edges. Let agents learn through play.

Success metric: Beat a baseline agent

Your Idea

Have something else? Pitch it. If it maps to a graph and benefits from collective optimization, it works.

What You’ll Build

1. TypeDB Schema

Define your domain concepts:

define

# Your entities
product sub entity,
  owns sku,
  owns name;

warehouse sub entity,
  owns location;

# Your relations
supply_route sub relation,
  relates origin,
  relates destination,
  owns distance,
  owns pheromone_level;

# Connect to colony infrastructure
warehouse plays supply_route:origin;
warehouse plays supply_route:destination;

2. Mission Configuration

mission:
  id: my-supply-chain
  name: "Supply Chain Optimization"

domain:
  schema: schema.tql

castes:
  scout: 0.30      # Explore new routes
  harvester: 0.50  # Exploit known routes
  relay: 0.20      # Spread information

success:
  metric: total_cost_reduction
  target: 0.10  # 10% reduction

3. Agent Behaviors

Customize how ants interact with your domain:

class SupplyChainScout(Scout):
    async def evaluate_edge(self, edge: Edge) -> float:
        # Custom success criteria
        delivery_time = await self.simulate_delivery(edge)
        return 1.0 / delivery_time  # Lower time = higher value

    async def deposit_pheromone(self, path: list[Edge], value: float):
        # Custom deposit logic
        for edge in path:
            await edge.add_pheromone(value * self.deposit_rate)

4. Visualization

Show your mission working:

• Pheromone evolution over time
• Agent allocation
• Performance metrics
• Superhighway formation

Infrastructure Provided

TypeDB Cloud

• Database: ants-colony
• Your mission gets isolated namespace
• Full TypeQL query access
• Schema write permissions

Agentverse

• Deploy up to 100 agents per team
• Run 24/7 during hackathon
• Real-time monitoring
• Logs and metrics

Compute

• GPU credits for heavy operations
• Background workers for simulation
• API access to colony services

Templates

• Starter mission template
• Example schemas
• Agent boilerplate
• Visualization components

Deliverables

1. Working Mission — Deployed and running on infrastructure
2. Schema — TypeDB schema for your domain
3. Agents — At least 2 caste types with custom behavior
4. Results — Measured performance against baseline
5. Demo — 5-minute live demonstration

Judging Criteria

The Prize

Winner gets:

• $5,000 cash prize
• Colony co-ownership (your mission becomes part of the permanent colony)
• Ongoing infrastructure access
• Potential for commercial development

Team Composition

Required:

• At least one developer (Python/TypeQL)
• At least one domain expert (for your chosen application)
• At least one non-CS person

Recommended:

• Backend developer (TypeDB, APIs)
• Frontend developer (visualization)
• Domain expert (supply chain, research, games, etc.)

Timeline

Getting Started

# Clone starter template
git clone https://github.com/antsatwork/mission-template

# Install dependencies
pip install -r requirements.txt

# Configure credentials
export TYPEDB_ADDRESS="https://cr0mc4-0.cluster.typedb.com:80"
export TYPEDB_DATABASE="ants-colony"

# Deploy your first scout
python deploy_scout.py --mission my-mission --caste scout --count 10

Mentors

• Tony O’Connell — ONE Ontology, can help with schema design
• Robin Dey — STAN algorithm, can help with agent behavior
• [DevRel] — TypeDB and Agentverse integration

“Talk is cheap. Show me the code.”

— Linus Torvalds

You’ve built apps, services, and systems.

Now build a colony.

[REGISTER FOR APPLICATION TRACK]