Director - AI Engineering

At Slalom, we co-create modern technology and software products with clients who are ready to accelerate AI-enabled digital product development. We imagine how things can be made better, then set out to realize what is possible - driving innovation with quality, resilience, and purpose. By blending design, product engineering, analytics, and automation, we build the AI-enabled custom software and data products of tomorrow.

As the Director of AI Systems and Platforms, you will manage a portfolio of client delivery and capability responsibilities while remaining an active technical leader. You will lead and develop AI/ML engineers, architects, and data scientists; shape and sell complex work; manage capability/practice health; and personally help teams design, build, and operate production AI systems. This role is for a leader who can move comfortably between client stakeholder conversations, architecture decisions, working sessions, code and design reviews, and production troubleshooting.

• Lead from the front on complex AI/ML engagements. Personally contribute to discovery, solution architecture, rapid prototypes, reference implementations, code and design reviews, performance tuning, and resolution of critical production issues.
• Architect, prototype, build, and product ionize agentic AI, retrieval-augmented generation (RAG), predictive ML, optimization, and intelligent workflow solutions. Translate business outcomes into technical designs, delivery increments, acceptance criteria, and measurable production results.
• Develop and review production code using Python, SQL, PySpark, and API frameworks such as FastAPI and Pydantic. Apply software engineering practices including automated testing, version control, secure coding, dependency management, and repeatable build and release processes.
• Build with vendor-balanced AI platforms and model ecosystems, including Microsoft Foundry and Azure Machine Learning;
  Amazon Bedrock and Amazon Sage Maker;
  Google Vertex AI and Gemini; and model or API ecosystems such as OpenAI, Anthropic, and Hugging Face.
• Apply agent and RAG frameworks such as Lang Graph/Lang Chain, Llama Index, and Semantic Kernel, together with vector and search technologies such as Azure AI Search, Open Search, Pinecone, Weaviate, pgvector, or equivalent services.
• Design and deliver data and AI workloads on Databricks and Snowflake, using technologies such as Spark, Delta Lake, MLflow, feature stores, model registries, and governed data products.
• Engineer containerized and cloud-native services using Docker, Kubernetes, managed compute, serverless patterns, event-driven integration, REST APIs, and streaming or batch pipelines.
• Establish MLOps and LLMOps practices across CI/CD, infrastructure as code, model and prompt versioning, automated evaluation, observability, lineage, security, and cost management. Use tools such as Git Hub Actions, Azure Dev Ops or Git Lab CI, Terraform, Open Telemetry, cloud-native monitoring, MLflow, and fit-for-purpose evaluation frameworks.
• Define and enforce production quality standards for groundedness, relevance, safety, latency, reliability, scalability, privacy, and responsible AI. Design human-in-the-loop controls, guardrails, auditability, and fallback patterns appropriate to each use case.

• Lead engagements from opportunity discovery and current-state assessment through architecture, delivery, production launch, and operational transition, ensuring alignment to business outcomes and technical excellence.
• Own solution and portfolio health across multiple work streams or engagements, including scope, staffing, delivery approach, dependencies, financial performance, risks, quality, and stakeholder communication.
• Create clarity across client and Slalom teams by defining architecture decisions, roles, delivery milestones, technical quality gates, and escalation paths.
• Step directly into delivery when risk is high or ambiguity is blocking progress - facilitating technical workshops, validating designs, pairing with engineers, reviewing implementation choices, or leading…