Madiff

Copilot Agent Engineer

Madiff  •  Remote  •  5 months ago
Apply
Save Job
Mark Applied Mark Interviewed Mark Offered
Hide Job Report Job
AI can make mistakes so check important info. Chat history is never stored.

Job Description


This is a remote position.


We are looking for a Copilot Agent Engineer to design the intelligence layer of an enterprise-grade AI assistant. The role focuses on shaping how the agent reasons, retrieves information, interprets intent, and operates across complex business workflows. This includes building structured reasoning frameworks, retrieval pipelines, evaluation layers, and ensuring enterprise-grade safety, auditability, and reliability.
​
Responsibilities

  • Design advanced prompt architectures and behavioural frameworks for Copilot


  • Build persona logic, agent roles, and structured conversational flows

  • Develop retrieval pipelines integrating Data Lake, Knowledge Graphs and internal APIs

  • Implement enterprise-grade RAG patterns with strong factual grounding and compliance

  • Create evaluation chains, deterministic rules and quality gates for sensitive workflows

  • Optimise agent outputs for accuracy, reliability, explainability, and traceability

  • Translate business policies and workflows into structured agent behaviour

  • Collaborate with engineering, platform, data, and business teams across workstreams

  • Ensure the Copilot meets strict audit, governance, and safety requirements

  • Build agent logic and orchestration flows using Python or TypeScript


    • Requirements


    • 4–6 years of experience in AI or backend engineering


  • Hands-on experience building
    LLM-powered agents
    (RAG, prompt engineering, orchestration)

  • Strong understanding of vector search, chunking strategies, embeddings and context optimisation

  • Ability to convert business logic into deterministic rules and reasoning structures

  • Experience with
    Copilot Studio, LangChain, Semantic Kernel
    or similar agent frameworks

  • Familiarity with prompt lifecycle management and output evaluation

  • Experience working in regulated environments where accuracy and compliance are critical

  • Strong technical experience with:

    • Azure OpenAI / GPT models

    • Vector databases (Pinecone, Weaviate, Chroma, Azure Cognitive Search)

    • Knowledge graphs and structured domain modelling

    • API orchestration and messaging layers

    • Proficiency in Python or TypeScript

    • Certifications such as:

    • Microsoft Azure AI Engineer / Azure Data Fundamentals

    • Generative AI or Applied AI Engineering foundations


    Nice to have
    ​


    • Azure OpenAI Specialisation

    • Exposure to knowledge engineering or ontology design

    • Additional cloud certifications

    • Experience with evaluation tooling (LangSmith, TruLens, custom eval frameworks)

    • Background in conversational orchestration or tool-calling architectures


    Benefits


    • Solid, competitive salary


  • Work in multinational environment on international projects

  • Comprehensive healthcare

  • Long-term B2B contract with stable project pipeline

  • Fully remote model
    • Madiff

    About Madiff

    We are an international Innovation, IT and high-tech engineering consulting company that delivers unique value in a wide variety of industries.

    Our mission is to add value to our customers businesses by providing digital and technological innovation services, delivering disruptive results and making our clients stand out in their market.

    We are driven by a creative and innovative consulting approach strongly oriented to getting results. We MAKE THE DIFFERENCE.

    Poland | UK | Switzerland | USA | Norway | Portugal | Spain | France | Singapore

    Warsaw

    Prosta 20

    Wrocław

    Rybacka 7

    Lublin

    Grottgera 2

    Industry
    IT & Software
    Company Size
    51-200 employees
    Headquarters
    Warszawa, PL
    Year Founded
    2015
    Website
    madiff.eu
    Social Media