The role
ResearchPricer is creating a machine-accessible route to new evidence: consumer surveys, expert interviews, focus groups, healthcare recruitment, B2B fieldwork and other custom research that must be collected after an agent asks for it. The engineering challenge is not simply model output. It is a governed transaction that crosses specifications, humans, suppliers, time zones, privacy requirements and money.
You will be the senior technical owner of that interface. You will work directly with the CEO, commercial team and Stratega CEE research operations to convert real quoting and delivery practices into secure product primitives. Early work is hands-on; architectural decisions must ship as reliable code, documentation, tests and observable production workflows.
What you will build
- Production MCP servers and agent-facing tools for capability discovery, brief validation, feasibility, quote preparation, order approval and project status.
- Secure REST APIs and structured contracts that remain predictable across models, clients and integration partners.
- Human-in-the-loop purchasing controls: organisation mandates, budget limits, approval policies, idempotency, audit trails and exception handling.
- Pricing and supplier-orchestration workflows spanning methodologies, audiences, countries, currencies, incidence assumptions and fieldwork risk.
- Payment and transaction foundations, including Stripe-style checkout, prepayment, webhook reconciliation, refunds and a clear boundary between quote and commitment.
- Operational interfaces that connect AI requests to Stratega CEE research teams, suppliers, respondent recruitment and structured delivery.
Your engineering remit
- Own architecture and implementation across Next.js, React, TypeScript, Vercel, Supabase and Postgres.
- Design authentication and authorisation for humans, organisations, service accounts and agents using least-privilege access.
- Model durable research and commerce state: briefs, feasibility versions, quotes, approvals, orders, payments, milestones, deliverables and cancellations.
- Build observability around latency, tool failures, model and provider costs, quote accuracy, webhook integrity and operational service levels.
- Create test harnesses for tool contracts, prompt-injection resistance, malformed agent requests, retries and concurrent transactions.
- Document interfaces for external developers and work directly with early integration partners through sandbox and production onboarding.
What we are looking for
- Strong production experience with TypeScript and modern server-rendered React or Next.js applications.
- Hands-on experience building LLM tools, agents, MCP servers or similarly structured machine-to-machine integrations.
- Deep understanding of API design, OAuth or service authentication, permissions, secrets, rate limits and secure webhook processing.
- Confident Postgres data modelling, migrations, indexes, row-level security and operational debugging; Supabase experience is valuable.
- Experience with payments, procurement, marketplaces or other workflows where idempotency, auditability and money movement matter.
- A practical approach to AI: deterministic boundaries, explicit status, measurable quality and human escalation where uncertainty is material.
- Clear written English and the ability to translate complex operational requirements into stable product contracts.
Especially useful
- Experience with Vercel Functions, durable workflows or queues, structured generation, OpenAPI and agent evaluation.
- Familiarity with market research, survey sampling, qualitative recruitment, healthcare research or multi-country fieldwork.
- Experience integrating Stripe, CRM systems, analytics, transactional email, storage and signed deliverable access.
- Security, privacy or compliance work involving GDPR, sensitive audiences, data minimisation or enterprise procurement.