Careers at Keele Labs
Build for the market's hardest moments.
Keele Labs is an early-stage team working at the intersection of on-chain execution, applied machine learning, and risk systems. We care about failure modes before demos and clear reasoning before speed.
Potential roles
Do exacting work with a small team.
These role areas are placeholders until hiring plans are confirmed. No location, compensation, or active-opening claims are implied.
MEV & on-chain execution
Design the transaction path from quote to confirmed execution across DEXs. Build for adversarial ordering, volatile liquidity, and observable failure.
- Build and test DEX adapters and routing systems
- Evaluate private order flow and MEV mitigations
- Own simulation, monitoring, and incident tooling
Strategy research & evaluation
Develop and challenge strategy hypotheses using realistic execution assumptions. Make costs, uncertainty, and failure cases visible.
- Design robust backtests and evaluations
- Model fees, slippage, liquidity, and MEV
- Define limits and live-monitoring criteria
Constrained AI agents
Build decision systems that turn market context into constrained, testable actions. Focus on evaluation, reproducibility, and safe fallback behavior.
- Develop agent policies and evaluation suites
- Convert model outputs into typed actions
- Measure errors, drift, and abstention behavior
Real-time infrastructure
Build the event-driven infrastructure behind market ingestion, state, execution, and monitoring.
- Operate reliable data and transaction pipelines
- Build replayable state and observability systems
- Design secure internal and external APIs
Key management & custody
Threat-model the systems that authorize and submit transactions. Make permissions narrow, visible, and revocable.
- Review signer and session-key architecture
- Assess contracts and third-party integrations
- Build incident response and disclosure processes
What matters here
Calm systems start with candid teams.
Own the failure mode
Ask how the system breaks before asking how well it demos.
Write the boundary
Turn assumptions into explicit interfaces, limits, and tests.
Change your mind
Disagree in specifics, update with evidence, and document the decision.