Vacancy details

Software Engineering

Python Engineer

Senior

Bulgaria, Croatia, Poland, Spain, Ukraine

Remote

APPLY NOW

REFER A FRIEND

You would help build a system that sits between complex map data and the teams that need usable simulation scenes. The work is partly product-facing and partly platform-facing: improving the search experience, strengthening the extraction pipeline, and making the output more reliable, scalable, and easy to consume.

This includes work on service integration, search logic, API-driven workflows, data transformation, and scene delivery. Depending on the role focus, it may also include improving performance, automation, output quality, and the flexibility of how scenes are generated and refined.

What project we have for you

Scene Extract is a platform designed to turn natural-language requests into usable driving scenes for simulation, testing, and analysis. Instead of requiring users to manually search through large map datasets and build scenes piece by piece, the platform lets them describe what they need in plain language and then automatically identifies, extracts, and prepares the relevant scene.
In simple terms, the platform helps users go from “describe the scenario” to “get a structured scene” much faster and with less manual effort.

The main purpose of Scene Extract project is to reduce the time and complexity involved in finding realistic road situations for development and validation work. It is especially valuable in environments where teams need many specific traffic, road, or infrastructure scenarios and cannot afford a slow, manual workflow.

The platform focuses on a few core outcomes:

Make scene discovery easier by allowing users to search with everyday language rather than technical queries.

Automate the extraction of relevant road and traffic context from available data sources.

Prepare scenes in a structured format that can be reused in downstream tools and workflows.

Improve consistency by standardizing how scenes are identified and generated.

At a high level, Scene Extract acts as a bridge between human intent and technical scene creation. A user provides a description of the situation they want, the platform interprets that request, locates matching map and environment elements, and produces an output that can be used for simulation or scenario-based work.

What you will do

Own full-stack architecture — make and document key technical decisions across frontend and backend, ensure they hold under scale and data volume
Define API contracts and frontend component architecture that the team builds against — clear, versioned, unambiguous
Lead technical discovery with BA/AI PM — translate product requirements into backend specs before they hit the engineers
Evaluate and decide on third-party integrations — Maps SDK, ArcGIS Enterprise, LLM providers
Set and enforce code quality standards across the backend — reviews, patterns, testing approach
Identify performance risks early — geospatial queries, OD matrix data volume, map rendering bottlenecks
Participate in client-facing technical discussions with the client’s PM and engineering team when needed
Mentor Engineers on backend concerns — keep the team moving without creating a bottleneck
*

What you need for this

Mandatory Requirements

5+ years of experience in software engineering with strong Python background
Excellent analytical, algorithmic and optimization skills,
OOAD, architecture and design patterns,
AWS — hands-on architecture experience (not just usage); able to design for performance and cost
TypeScript — advanced daily usage across frontend and backend
REST API design — versioning, contract-first approach, documentation
SQL — complex queries, query optimization, understanding of data modeling
Experience integrating third-party APIs — preferably data-heavy or geospatial
Hands-on experience with AI coding agents — Cursor, Claude Code, GitHub Copilot, Codex, or similar; part of your daily engineering workflow, not a side experiment
Proven experience in a tech lead or principal role — own decisions across the full stack, not just execute them
RAG system design and implementation — chunking strategies, embedding models (OpenAI, Cohere, or open-source), vector DB (pgvector, Pinecone, Weaviate, or similar), retrieval pipelines end to end
Agentic workflow implementation — tool use / function calling, multi-step reasoning chains, agent state management, LLM orchestration (LangChain, LlamaIndex, or custom)
Prompt engineering at system level — production system prompts, few-shot patterns, chain-of-thought for structured data queries, context window management
LLM evaluation — building evaluation pipelines: metrics, test sets, regression testing on model or prompt updates
English — B2+, comfortable in direct client communication and async written specs

Nice to Have

React – able to right frontend using React
Geospatial data processing — spatial datasets, coordinate systems, tiling pipelines, or aggregation formats (H3, S2, GeoJSON, WKT)
Map rendering experience — Maps SDK, Mapbox GL JS, deck.gl, or WebGL-based visualization; understanding of performance constraints at scale
MCP (Model Context Protocol) — experience implementing MCP server or client; understanding of how it fits into agentic tool ecosystems
Guardrails and safety layers — input/output validation for LLM, fallback logic, AI-specific rate limiting
Streaming inference and latency optimization — streaming API responses, token budgeting, caching strategies for LLM calls
LLM observability — LangSmith, Helicone, or custom tracing for prompt/response logging in production
Spec-driven development — contract-first API design, OpenAPI authoring, working from or producing formal functional specs
Public sector or traffic/mobility domain background
Experience shipping SaaS products with tiered access / feature flagging
Familiarity with data pipeline design — even if not owning the computation side

AI-Driven Development — How We Work

This team operates as an AI pod — AI tooling is a core part of the engineering workflow at every level. As engineer, you are expected not just to use AI tools yourself, but to actively shape how the team adopts and integrates them into the development process.

We are looking for someone who has a strong opinion on AI-assisted engineering and can lead by example — helping a lean team punch above its weight.

Deep daily use of AI coding assistants — Cursor, GitHub Copilot, or equivalent; strong personal workflow already in place
Experience evaluating and adopting new AI tooling — you follow the space and have opinions on what works
Comfortable with agentic development workflows — Claude Code, Devin, or similar; understands where to trust and where to verify
Experience using LLMs for architecture exploration, spec drafting, code review assistance — beyond just code generation
Able to define AI tooling standards and best practices for the team — what to use, how to review AI output, where the guardrails are

What it’s like to work at Intellias

At Intellias, where technology takes center stage, people always come before processes. By creating a comfortable atmosphere in our team, we empower individuals to unlock their true potential and achieve extraordinary results. That’s why we offer a range of benefits that support your well-being and charge your professional growth.
We are committed to fostering equity, diversity, and inclusion as an equal opportunity employer. All applicants will be considered for employment without discrimination based on race, color, religion, age, gender, nationality, disability, sexual orientation, gender identity or expression, veteran status, or any other characteristic protected by applicable law.
We welcome and celebrate the uniqueness of every individual. Join Intellias for a career where your perspectives and contributions are vital to our shared success.

Senior Python Engineer