Mid-Level AI Engineer

About the role

Dogma Group is looking for a Mid-Level AI Engineer to join its client services team. You will design and deliver AI‑driven solutions under tight deadlines, focusing on agentic systems powered by large language models and computer‑vision applications.

Key responsibilities

• Design, build, and deploy single‑agent and multi‑agent systems using LangChain, LangGraph or custom Python code with Pydantic.
• Connect agents to tools, data sources, and other agents via standardized protocols such as MCP and A2A.
• Develop production‑grade Retrieval‑Augmented Generation (RAG) pipelines with hybrid dense‑vector and keyword search, and rerank results using cross‑encoders.
• Choose and integrate vector databases (Pinecone, Weaviate, Qdrant, pgvector) based on client scale and data needs.
• Implement computer‑vision solutions (object detection, classification, segmentation, OCR) using OpenCV, TensorFlow or PyTorch, and vision‑language models.
• Containerize services with Docker and manage deployments across cloud and on‑premise environments.
• Maintain observability, tracing, and evaluation pipelines to monitor model performance, drift, and reliability.
• Translate client requirements into technical scope, build data pipelines, and deliver iteratively.
• Participate in client calls, demos, and technical reviews.

Required profile

• 3+ years of hands‑on AI/ML development and deployment experience.
• Bachelor’s degree in Computer Science, Engineering, Data Science, or a related field.
• Strong foundation in machine‑learning, deep‑learning, statistics, probability, linear algebra, and calculus.
• Proven experience with generative AI, large language models, and agentic architectures.

Required skills

• Python programming
• LangChain / LangGraph
• Pydantic
• Model Context Protocol (MCP) and Agent‑to‑Agent (A2A) protocols
• RAG pipeline design
• Vector databases: Pinecone, Weaviate, Qdrant, pgvector
• OpenCV, TensorFlow, PyTorch
• Docker containerization
• Version control (Git) and testing practices