Principal AI Engineer

Requirements

• Designing and building AI-powered systems in production
• Agentic frameworks, multi-agent collaboration, orchestrator/worker models
• RAG pipelines, relevance tuning
• Python and/or TypeScript, API design, microservices, cloud-native architectures
• Presales consulting, translating business problems into AI solutions

Additional requirements:

• Practical knowledge of multiple LLM providers (OpenAI, Anthropic, open-source)
• Ability to build/adapt models (prompt-tuned agents, LoRA fine-tuning, inheritance from foundation models)
• Knowledge of AWS Bedrock, Azure OpenAI, GCP Vertex AI
• Understanding governance, security, data residency, pricing models, enterprise integration for cloud AI platforms
• Production-grade mindset: observability, logging, security, PII handling, cost-efficiency
• Strong communication skills for explaining AI concepts to non-technical stakeholders
• English level: Upper-Intermediate
• Experience: 5+ years in AI engineering or related roles

WILL BE A PLUS

• Experience in AdTech domain
• Experience integrating AI into legacy or enterprise systems
• Familiarity with workflow automation tools
• Knowledge of DevOps practices, CI/CD, containerization

Responsibilities

• Consult clients during presales to assess AI readiness, constraints, and success criteria, translating visions into actionable requirements
• Prepare and present presales deliverables including architecture diagrams, assumptions, risks, estimates, scalability considerations, and implementation roadmaps
• Define AI use cases and agentic scenarios based on client needs
• Select and justify LLMs per use case based on requirements, cost, latency, safety, capability
• Architect multi-agent frameworks with orchestrator agents and MCP-style protocols
• Design and implement single-agent/multi-agent AI systems with defined roles, tool access, memory, safety boundaries
• Build orchestration logic (routing, delegation, retries, fallback strategies, consensus, human-in-the-loop flows)
• Develop RAG pipelines (data ingestion, chunking, embeddings, vector databases, hybrid retrieval, relevance optimization)
• Implement learning & feedback loops for continuous agent improvement
• Design custom or adapted models (prompt-tuned agents, LoRA fine-tuning, domain-inherited models)