AI Solutions Architect (Freelance)

We are looking for an experienced AI Solutions Architect to lead the design and implementation of AI-powered solutions across our product and engineering landscape. This role is instrumental in shaping our AI architecture strategy, driving automation with intelligent agent systems, and modernising software engineering workflows using the latest advancements in AI tooling, protocols, and frameworks.

As AI moves from experimentation to enterprise-scale production, you will be at the forefront — designing systems that are scalable, observable, and governed. You will bridge business strategy and deep technical execution, while championing responsible and secure AI practices.

Innovecs is an equal opportunity employer. All hiring decisions are based on professional qualifications, skills, and experience. We are committed to a transparent, merit-based recruitment process that prevents discrimination and ensures equal opportunities for all candidates. Reasonable accommodations are available upon request throughout the recruitment process to support accessibility and inclusion.

Technology Stack

* LLM & Agent Frameworks: LangChain, LangGraph, LangWatch/CrewAI, AutoGen, MS Copilot Studio, OpenAI Agents SDK, Anthropic Claude SDK, Mastra, etc.
* Languages: Python, Go, TypeScript, etc.
* AI/ML Frameworks: PyTorch, TensorFlow, scikit-learn, ONNX, Hugging Face Transformers;
* Agent Protocols: MCP, A2A, ACP;
* Orchestration: Services and tools for multi-agent-pipelines, observability, agent lifecycle and registry management;
* Model Serving & APIs: vLLM, TorchServe, TensorFlow Serving, FastAPI, gRPC
* Vector Databases: Pinecone, Weaviate, pgvector, Chroma, Qdrant. etc.
* Data & Streaming: Apache Spark, Kafka, Pandas, NumPy, dbt;
* MLOps & Experiment Tracking: MLflow, Kubeflow, Airflow, Metaflow, DVC, Weights & Biases;
* Feature Stores: Feast, Tecton;
* Cloud AI Platforms: Google Vertex AI, AWS SageMaker, Azure ML, AWS Bedrock, GCP AI Platform;
* Infra & Orchestration: Kubernetes, Docker, Helm, Terraform;
* Observability & Monitoring: Prometheus, Grafana, ELK Stack, OpenTelemetry, Arize AI;
* CI/CD & Version Control: GitHub Actions, Jenkins, Git, ArgoCD;
* AI Dev Tooling: Claude Code, Cursor, GitHub Copilot, Windsurf;
* Security & Governance: OAuth 2.x, RBAC, audit logging for MCP, AI model governance frameworks;
* Compliance: ISO 27001, SOC2, EU AI Act, EU Data Act, CRA, GDPR.

Position requirements

* Must-Have:
* 5+ years of experience in AI/ML solution architecture and demonstrable track record of taking AI systems from prototype to production at scale;
* Deep expertise in LLMs, prompt and context engineering, RAG architectures, and vector databases;
* Hands-on experience with agentic AI frameworks and orchestration;
* LangChain / LangGraph: multi-step reasoning chains and stateful agent workflows;
* LangWatch / CrewAI / AutoGen: multi-agent collaboration and task delegation;
* MCP (Model Context Protocol): designing and deploying MCP servers for agent-to-tool integration;
* Strong proficiency in Python; working knowledge of at least one additional language (Go, TypeScript, etc.);
* Experience with cloud-native AI deployment on AWS, GCP, Azure, including managed LLM services such as Bedrock, Vertex AI, etc.
* Solid software engineering fundamentals: design patterns, API design, testing, and CI/CD;
* Familiarity with AI observability, evaluation frameworks, and production monitoring of LLM-based systems;
* Ability to define AI adoption roadmap, prioritize business cases based on ROI, present the strategic and tactical layers of implementation to both technical and business stakeholders;
* Experience with AI compliance, governance frameworks, and explainability (GDPR, EU AI Act, model cards);
* Experience integrating AI into enterprise systems (ERP, CRM, ITSM) via standardised protocols;
* Experience leading engineering teams on AI-first projects.
* Nice-to-Have:
* Background in AI security: prompt injection mitigation, MCP server hardening, OAuth 2.x for agents;
* Knowledge of emerging multi-agent communication standards: A2A (Google), ACP (IBM BeeAI);
* Experience with reasoning/thinking models and their architectural implications for agent planning;
* Contributions to open-source AI projects, MCP servers, or published technical articles.

Position responsibilities

* AI Products & Solution Architecture:
* Design and guide implementation of AI-driven products, APIs, and platform features from concept to production;
* Evaluate, select, and benchmark AI/ML models — including frontier LLMs, fine-tuned models, and open-source alternatives;
* Architect scalable, observable, and cost-efficient AI systems that span experimentation, staging, and production;
* Collaborate with product managers and business stakeholders to translate requirements into robust solution architectures;
* Establish architectural standards for multi-agent systems, including context management strategies and memory designs.
* Agentic AI & Process Automation:
* Identify business processes that can be automated or enhanced via agentic AI, and define the architecture for doing so;
* Design and oversee implementation of MCP server ecosystems that connect agents to enterprise data sources and tools;
* Architect multi-agent workflows using orchestration frameworks (LangGraph, CrewAI, AutoGen), with appropriate human-in-the-loop checkpoints;
* Integrate agent-to-agent communication standards (A2A, ACP) where multi-agent coordination is required;
* Drive governance of MCP deployments: audit trails, authentication, rate limiting, and access control policies;
* Embed AI into internal and external tools to improve operational efficiency across teams.
* AI-Augmented Software Engineering:
* Set up and continuously optimize AI-augmented developer environments (Claude Code, Cursor, GitHub Copilot);
* Introduce AI into automated testing, deployment pipelines, code review, estimation, and technical documentation;
* Define and enforce best practices for using AI coding tools safely, securely, and productively in software delivery;
* Drive adoption of context engineering disciplines — designing prompts, tool schemas, and MCP resources that maximize agent reliability;
* Governance, Security & Responsible AI:
* Ensure all AI systems are designed with security-first principles: input validation, output guardrails, and least-privilege access;
* Maintain AI compliance standards aligned with GDPR, the EU AI Act, EU Data Act, CRA, ISO27001, and internal model governance policies;
* Implement observability and evaluation pipelines to detect hallucinations, drift, and performance degradation in production LLM systems.