AI Engineer

Level

Middle

Department

Type

Full Time

Project

NDA

Locations:

Romania

Job Details

Posted on:

September 1, 2026

Job ID:

YzgNuz

About the Company

Established in 2004, ALLSTARSIT was founded with a clear vision: to enhance the landscape of global IT employment by bridging the gap between companies and skilled professionals. The core belief was that assembling a team shouldn't be hindered by geographical constraints. Fast forward to the present day, ALLSTARSIT stands as an international outstaffing service provider committed to change the way businesses recruit, compensate, and oversee top talent worldwide.

With operational hubs scattered across Europe, Asia, and LATAM, and its headquarters situated in San Francisco, US, the company boasts a workforce of over 1,000 adept professionals. Spanning across more than 20 countries, ALLSTARSIT offers a diverse range of skilled employees across various verticals, including AI, cybersecurity, healthcare, fintech, telecom, media, and so on.

About the Project

Our client is an applied AI lab building semantic AI that can actually think with you — grounded in formal knowledge representations and sound reasoning. Their mission is to create trusted agents that reason over deep, structured knowledge: texts, commentaries, arguments, and traditions. We’re helping them look for an AI Engineer to build the LLM layer of our platform: multi-agent workflows, hybrid retrieval over knowledge graphs and vector indexes, inference integration, and evaluation. You’ll sit between research and platform — taking agent architectures from prototype to production and making them measurably reliable.

Specialization

Headquarters

Years on the market

Team size and structure

Current technology stack

Required skills:

Must Have

4+ years of software engineering experience (backend or ML), including production systems in Python.
Hands-on experience building LLM systems beyond demos: agents and tool use, RAG, or evaluation pipelines.
Real workflow experience with the OpenAI/Anthropic APIs (or comparable).
Solid engineering fundamentals: API design, services, testing, deployment.

Nice to Have

Structured knowledge representations: ontologies, knowledge graphs, SPARQL, or graph databases (e.g., Neo4j).
Vector databases and hybrid retrieval architectures.
Kubernetes and cloud-native deployment.
Model serving (e.g., vLLM), fine-tuning, or evaluation frameworks.
Go and/or Scala.

Scope of work:

Build and productionize multi-agent workflows on Anthropic/OpenAI APIs: orchestration, tool use, structured outputs, guardrails.
Design hybrid retrieval architectures that combine knowledge graphs, vector search, and ranking into a single coherent context layer.
Build evaluation harnesses and observability for agent behavior — quality, latency, cost — and use them to drive iteration.
Integrate LLM inference, retrieval, and reasoning services into production backends.
Work with researchers and domain experts to turn neuro-symbolic prototypes into robust product features.