Product Engineer | Callimacus.ai

1 mese fa

Lombardia, [null, null], Italia Welyk Tempo pieno 45.000 € - 85.000 € Contratto

Way of work: On site Milano, zona Moscova
Key tech skills:
TypeScript (Node.js), React, AWS, RAG/LLM/agent orchestration
Experience:
Middle/Senior
Total comp:
45-85k (45-60k middle, 70-85k senior) plus equity and benefits

About Callimacus

Callimacus is reinventing how people interact with the web. Since the Nineties, digital experiences have been built around a fixed paradigm: predefined pages, menus, and navigation trees, that no longer fits how people browse, shop, and get informed today.

Callimacus built the engine for something different: a multi-agentic platform that plugs into a company's existing stack, reads real-time signals to understand what a visitor wants, and composes the entire interface around that intent in milliseconds, selecting the most relevant content and components as the person interacts.

Founded in 2024 with early backing from Brunello Cucinelli, now both an investor and a customer, Callimacus launched in January 2026 to strong enterprise interest, and recently raised an important investment from Salesforce Ventures, its largest AI investment in Italy. The team is small, product-focused, and founder-led, working on hard problems nobody has solved before.

What you'll build

You'll join a small senior team at the point where the platform's foundations are still being set. Callimacus is in production with demanding customers, and the architecture, the performance envelope, and the shape of the developer experience are all still yours to influence. Decisions you make in your first months will still be load-bearing in three years.

  • Own the orchestration layer. Design how our agents turn multi-dimensional visitor signals into intent, and intent into a fully composed interface — in milliseconds, in production, for demanding enterprise customers.

  • Put inference on the critical path — and keep it fast. Model routing, caching, fallbacks, and latency budgets for a system where “the model is thinking” is not an acceptable state.

  • Build the retrieval and selection pipeline. Embeddings, vector search, ranking, and the content-selection logic that decides what a visitor sees next.

  • Make agent behavior measurable. Evals, tracing, and feedback loops, so quality is an engineering discipline rather than a feeling.

  • Shape how customers extend the system. Tools, guardrails, and configuration — the primitives others will build on.

One note before applying

Callimacus believes working side by side leads to faster progress, stronger ramp-up, and a culture built around community.
For the next months, the team has chosen to be in-person first, they've invested in making the office worth coming to, and are happy to support your commute.

They also don't work by tickets: priorities get redefined often, sometimes mid-week, and you'll need to be comfortable with that. If you're set on remote work or a settled roadmap, this isn't your role. If you want to be in the room while the platform takes shape, they'd like to meet you.

What you'll build

You'll join a small senior team at the point where the platform's foundations are still being set. Callimacus is in production with demanding customers, and the architecture, the performance envelope, and the shape of the developer experience are all still yours to influence. Decisions you make in your first months will still be load-bearing in three years.

  • Own the orchestration layer. Design how the agents turn multi-dimensional visitor signals into intent, and intent into a fully composed interface in milliseconds, in production, for demanding enterprise customers.

  • Put inference on the critical path, and keep it fast. Model routing, caching, fallbacks, and latency budgets for a system where "the model is thinking" is not an acceptable state.

  • Build the retrieval and selection pipeline. Embeddings, vector search, ranking, and the content-selection logic that decides what a visitor sees next.

  • Make agent behavior measurable. Evals, tracing, and feedback loops, so quality is an engineering discipline rather than a feeling.

  • Shape how customers extend the system. Tools, guardrails, and configuration, the primitives others will build on.