We're looking for an AI Specialist to expand our client`s team in Lviv! This is a full-time, on-site position with a leadership role in applied research and developing open-source LLM models for a large enterprise customer. You'll integrate and optimise models deployed via Ollama, pushing the boundaries of enterprise AI use cases.
Responsibilities:
- Shape our GenAI & classical ML vision.
- Advise on build-vs-buy for LLM stacks, vector DBs, and MLOps tooling.
- Address similarity & normalisation
- Design deep-learning + rule pipelines to match ~250 M address records/year.
- Evaluate with ragas and bespoke geo-accuracy metrics.
- Customer predictions
- Engineer features from parcel scans, seasonal peaks, traffic data.
- Train and ship predictive models (ETA, delivery-success likelihood) instrumented with Langfuse.
- Semantic search & RAG
- Implement RAG on FAQs, SOPs, and support tickets using open-source LLMs (e.g., Mixtral, Phi-3) served via Ollama.
- Optimise embedding stores, latency, and cost for call-centre workloads.
- Production excellence
- Deploy models to Kubernetes/AWS with CI/CD, monitoring, canary roll-outs.
Requirements:
- Open-source LLMs: Familiarity with open-source LLMs (LLaMA, Mistral, Phi, etc.).
- Deployment Tools: Experience using or deploying via Ollama, Hugging Face, or similar tools.
- Independence & Communication: Ability to work independently and communicate clearly in English.
- Research Mindset: Strong research mindset, hands-on coding ability, and documentation skills.
What We Offer:
- Leadership Role: A leadership role in an applied AI initiative for an international enterprise.
- Cutting-edge Tools: Work with cutting-edge tools like Ollama, RAG pipelines, and low-latency deployments.
- Talented Team: A talented, motivated team.
- Competitive Compensation: Competitive compensation and long-term opportunity.
Location: Lviv, Ukraine | On-site, Full-time
