Senior AI Engineer (LLM / Multi-Agent Systems)
OnHires • Portugal
Posted: May 4, 2026
Job Description
Senior AI Engineer (LLM / Multi-Agent Systems)
Remote (EU) · Full-time · Core product role
About the product
We’re building DOGER — a production system that connects real-world public data (real estate, weather, statistics, fire events, tenders) into a unified graph and answers cross-domain questions.
Think:
“Is this property a good investment based on tourism trends, fire risk, and pricing?”
“Are there anomalies in public tenders after major events?”
The system is already live and growing fast.
New data sources and use cases are added every 1–2 weeks.
The role
You’ll own the AI layer of the product.
Everything between:
👉 “user asks a question”
👉 “system returns a structured, reliable answer”
This includes how data is retrieved, combined, reasoned over, and turned into outputs.
You won’t be starting from scratch — but you will take ownership and evolve a working system.
What you’ll work on (next 3–6 months)
1. Expand system capabilities
add new use cases and data sources (weekly / bi-weekly)
extend how the system reasons across domains
improve answer quality and structure
2. Orchestration and system logic
improve multi-step workflows (agents, tools, routing)
design how different parts of the system interact
make behavior more predictable and debuggable
3. External access layer
enable external systems to query the platform
build APIs and access patterns for data and reasoning
prepare for monetization (e.g. per-query access, integrations)
4. Understand and improve the current system
dive into an existing system that’s partially a “black box”
map how it works end-to-end
refactor where needed (without full rewrites)
What already exists
production LLM-powered system
graph-based data layer (Neo4j)
partially automated agent / workflow creation (~80%)
multiple real-world data sources connected
ability to add new data sources in 1–2 days
working UI and real use cases
This is not a prototype — it’s a system already delivering value.
Tech stack
Core:
Python
LLM frameworks (Pydantic AI, LangChain, LangGraph, OpenAI SDK, or similar)
APIs and data pipelines
Nice to have:
graph databases (Neo4j or similar)
FastAPI or backend frameworks
experience with multi-step LLM workflows (agents, tool use, orchestration)
Who we’re looking for
You’re likely a strong fit if you:
have built LLM-powered systems in production (not just demos)
understand how to structure AI systems (RAG, tools, workflows, APIs)
can debug and improve non-deterministic behavior
are comfortable working with messy, evolving systems
have worked in startups or high-ownership environments
What matters most
ownership mindset
ability to move fast and iterate
ability to understand and improve existing systems
strong engineering fundamentals
Team
Small, product-focused team (currently 4 people):
Product / project lead
Data & backend engineer (data pipelines, infra)
DevOps
You — owning the AI layer
Working format
full-time (no part-time)
remote
overlap with Portugal working hours for collaboration
Location
Priority:
Portugal (RU/UA speakers preferred)
Portugal (English-speaking)
Europe (English / RU / UA)
Flexible for the right person.
Contract
B2B contract (Dubai entity)
compensation discussed individually
Why this role
real product, not a prototype
direct impact on what gets built and shipped
ownership of a core system, not a small feature
fast iteration, minimal process overhead
Important to know
This is a high-ownership role:
you’ll be the main person responsible for the AI system
things move fast
not everything is perfectly structured yet
If you enjoy building, improving, and owning systems — this will fit.
If you prefer clearly defined boundaries — probably not.
Additional Content
Senior AI Engineer (LLM / Multi-Agent Systems)
Remote (EU) · Full-time · Core product role
About the product
We’re building DOGER — a production system that connects real-world public data (real estate, weather, statistics, fire events, tenders) into a unified graph and answers cross-domain questions.
Think:
“Is this property a good investment based on tourism trends, fire risk, and pricing?”
“Are there anomalies in public tenders after major events?”
The system is already live and growing fast.
New data sources and use cases are added every 1–2 weeks.
The role
You’ll own the AI layer of the product.
Everything between:
👉 “user asks a question”
👉 “system returns a structured, reliable answer”
This includes how data is retrieved, combined, reasoned over, and turned into outputs.
You won’t be starting from scratch — but you will take ownership and evolve a working system.
What you’ll work on (next 3–6 months)
1. Expand system capabilities
add new use cases and data sources (weekly / bi-weekly)
extend how the system reasons across domains
improve answer quality and structure
2. Orchestration and system logic
improve multi-step workflows (agents, tools, routing)
design how different parts of the system interact
make behavior more predictable and debuggable
3. External access layer
enable external systems to query the platform
build APIs and access patterns for data and reasoning
prepare for monetization (e.g. per-query access, integrations)
4. Understand and improve the current system
dive into an existing system that’s partially a “black box”
map how it works end-to-end
refactor where needed (without full rewrites)
What already exists
production LLM-powered system
graph-based data layer (Neo4j)
partially automated agent / workflow creation (~80%)
multiple real-world data sources connected
ability to add new data sources in 1–2 days
working UI and real use cases
This is not a prototype — it’s a system already delivering value.
Tech stack
Core:
Python
LLM frameworks (Pydantic AI, LangChain, LangGraph, OpenAI SDK, or similar)
APIs and data pipelines
Nice to have:
graph databases (Neo4j or similar)
FastAPI or backend frameworks
experience with multi-step LLM workflows (agents, tool use, orchestration)
Who we’re looking for
You’re likely a strong fit if you:
have built LLM-powered systems in production (not just demos)
understand how to structure AI systems (RAG, tools, workflows, APIs)
can debug and improve non-deterministic behavior
are comfortable working with messy, evolving systems
have worked in startups or high-ownership environments
What matters most
ownership mindset
ability to move fast and iterate
ability to understand and improve existing systems
strong engineering fundamentals
Team
Small, product-focused team (currently 4 people):
Product / project lead
Data & backend engineer (data pipelines, infra)
DevOps
You — owning the AI layer
Working format
full-time (no part-time)
remote
overlap with Portugal working hours for collaboration
Location
Priority:
Portugal (RU/UA speakers preferred)
Portugal (English-speaking)
Europe (English / RU / UA)
Flexible for the right person.
Contract
B2B contract (Dubai entity)
compensation discussed individually
Why this role
real product, not a prototype
direct impact on what gets built and shipped
ownership of a core system, not a small feature
fast iteration, minimal process overhead
Important to know
This is a high-ownership role:
you’ll be the main person responsible for the AI system
things move fast
not everything is perfectly structured yet
If you enjoy building, improving, and owning systems — this will fit.
If you prefer clearly defined boundaries — probably not.