AI Agent Engineer (80,000 - 140,000 MDL)

Role Overview:
As our AI Agent Engineer, you’ll be the architect behind the “smarts” of Yolk’s platform. You’ll design and build modular LLM-p
- Analyze sales interactions to identify skill and behavior gaps
- Generate targeted scenarios grounded in best-practice examples
- Benchmark & optimize LLMs (e.g. Claude vs. GPT) for accuracy, cost, and token efficiency
- Deliver structured insights into our platform’s dashboards for rapid action

What You’ll Do:
- Agent Orchestration & Prompt Engineering
Architect multi-step pipelines (e.g. gap extraction → scenario generation → evaluation) using LangChain or equivalent.
- Data Pipeline Development
Build scalable flows to preprocess and index sales interaction data; implement RAG strategies (LlamaIndex, Haystack, etc.) to gr
- Model Evaluation & Comparison
Define metrics (BERTScore, ROUGE, cost/token) and A/B frameworks; generate dashboards and reports that guide model select
- Token & Cost Optimization
Profile end-to-end context lengths; prototype summarization, caching, or retrieval-first strategies to minimize spend.
- Light Fine-Tuning & Adaptation
Apply adapter-based methods (LoRA) or API-level tuning to hone models on sales-specific language.
- Collaboration & Handoff
Author Dockerfiles or Kubernetes manifests for your services; partner with DevOps on deployment and with front-end on API co

Must-Have Qualifications:
- Bachelor’s in Computer Science, AI/ML, or equivalent practical experience
- 3+ years of professional Python development (async, testing, packaging)
- Deep hands-on experience with LangChain (or comparable agent orchestration frameworks)
- Proven work with dense embeddings & vector search (FAISS, Pinecone, Chroma)
- Proficiency in NLP evaluation libraries (Evaluate, rouge-score, bert-score)
- Solid understanding of RAG concepts and tools (LlamaIndex, Haystack)
- Experience with OpenAI/Anthropic APIs, including function-calling and rate-limit strategies
- Ability to write Dockerfiles/Kubernetes manifests—DevOps handles cluster ops

Nice-to-Have & Bonuses:
- Familiarity with Microsoft Autogen or similar advanced agent frameworks
- Experience serving models via BentoML or Seldon Core
- Knowledge of monitoring & experiment tracking tools (LangSmith, MLflow, Evidently AI)
- Background in sales operations, conversational UX, or competitive intelligence
- Exposure to lightweight front-end frameworks (Streamlit, basic React)

Why You’ll Love working with us:
- High Impact: Shape the core AI that powers our flagship Sales OS.
- Deep-Tech Backing: Funded by the same investors behind Anthropic, TensorNet, and GROQ.
- Very competitive Salary from 4,000 Euro till 7,000 Euro depending on the experience and interview.
- Collaborative Culture: Partner with world-class DevOps, back-end, and front-end teams.
- Growth Opportunity: Leverage our research group for advanced model development down the road.