THE ROLE
Process & Agent Design
Work directly with the founder to understand how customers operate and where agents can help. Your focus is discovery and design: uncover the real problem, make processes and business knowledge explicit, and translate them into useful agent workflows. Build practical mapping tools and make focused code changes when they help move the solution forward.
Dialetica builds and operates AI agents that work alongside humans inside companies. We are an early team, with open problems and room to shape both the product and how the company works.
What you will do
- Understand the customer’s pain. Talk to users, observe their work and distinguish symptoms from root causes. Identify where time is lost, decisions get stuck or existing tools fail, and define what a useful improvement would look like.
- Map processes and agent flows. Document activities, decisions, systems and exceptions. Understand how an agent receives context, uses tools and passes work to other agents or humans, with clear ownership at each step.
- Build tools for ontology mapping. Create practical ways to capture and explore the customer’s business entities, relationships, terminology and rules. Turn that shared understanding into structured knowledge that people and agents can use.
- Make targeted implementations. Use small scripts, lightweight tools and focused integration or configuration changes to validate a workflow or remove a specific bottleneck. Work with the engineering-focused teammate when deeper implementation is needed.
- Validate and refine with users. Test whether the mapped workflow reflects reality and solves the intended problem. Study agent behavior and handoffs, update the design with evidence and translate recurring needs into product improvements.
What you bring
- You enjoy listening to customers, asking precise questions and turning ambiguous needs into clear process maps.
- You are curious about how a business represents its knowledge: entities, relationships, rules and the context an agent needs to act.
- You can move between discovery and hands-on work, using code for focused problems without making full-time software development the center of the role.
What we look for
- Practical software problem-solving. Some programming knowledge, or demonstrated ability to use ClaudeCode/Codex/Cursor to build and debug solutions. You should be able to explain what you built, understand its limits and verify that it works.
- Scientific curiosity about agentic systems. An interest in planning, tool use, memory, evaluation and multi-agent coordination. Read, form hypotheses, run experiments and update your views when the evidence changes.
- Clear communication and ownership. Listen carefully, ask useful questions, make trade-offs explicit and follow work through to a verified outcome.
- Evidence of learning and execution. A project, integration, analysis, automation or problem you solved. Work, study and personal projects all count.
SHOW US YOUR WORK
How you think.
What you have built.
Tell us about your experience, something you built or a problem you solved, and a question about AI agents you would like to investigate. A repository, demo or clear account of your work is a good place to start.