AI Lead - Platform Intelligence & Applied AI-REMOTE
Job Description
AI Lead – Platform Intelligence & Applied AI-REMOTE
\n
\n
Core Responsibilities
\n
AI Strategy & Market Intelligence
\n
• Continuously track and evaluate:
\n
o LLM and foundation model advancements
\n
o Agent frameworks and orchestration patterns
\n
o Retrieval, memory, and context management techniques
\n
o AI evaluation, safety, and governance approaches
\n
• Translate emerging AI trends into:
\n
o Platform design principles
\n
o Proofs of concept and experiments
\n
o Scalable, production-ready capabilities
\n
• Advise leadership on when and how new AI capabilities should be adopted.
\n
Model & Intelligence Management
\n
• Own the strategy for LLM and model usage across the platform, including:
\n
o Model selection and benchmarking
\n
o Versioning and lifecycle management
\n
o Cost, performance, and latency trade-offs
\n
o Fallback and redundancy strategies
\n
• Establish best practices for:
\n
o Prompt and instruction design
\n
o Skill and Tool calling
\n
o Structured outputs and determinism
\n
Semantic Routing & Orchestration
\n
• Design and evolve the platform’s semantic routing layer, including:
\n
o Intent detection and task classification
\n
o Routing to appropriate models, agents, or workflows
\n
o Context-aware decisioning based on workspace state
\n
• Define orchestration patterns for:
\n
o Multi-step and parallel execution
\n
o Long-running and asynchronous tasks
\n
o Human-in-the-loop controls
\n
• Ensure routing logic is transparent, testable, and tunable.
\n
Agent Architecture & Execution
\n
• Consult on the firm’s agent strategy, including:
\n
o When to use agents vs. workflows vs. direct LLM calls
\n
o Agent composition, memory, and tool access
\n
o Guardrails and behavioral constraints
\n
• Partner with engineering to implement:
\n
o Agent frameworks and runtime infrastructure
\n
o Monitoring and debugging capabilities
\n
• Ensure agents are:
\n
o Predictable and auditable
\n
o Aligned to service methods and delivery workflows
\n
o Safe for enterprise and client-facing use
\n
Workspace Context & RAG Architecture
\n
• Own the design of contextual intelligence within workspaces, including:
\n
o Document ingestion, chunking, and enrichment strategies
\n
o Vector, keyword, and hybrid retrieval approaches
\n
o Context assembly across client data, firm IP, and engagement artifacts
\n
• Define standards for:
\n
o Source attribution and transparency
\n
o Data isolation and compliance
\n
o Relevance, freshness, and performance
\n
• Continuously evaluate new approaches to memory, retrieval, and grounding.
\n
AI Evaluation, Testing & Trust
\n
• Establish the platform’s AI evaluation and testing framework, including:
\n
o Task-based and scenario-driven evaluations
\n
o Regression testing for prompts, agents, and routing logic
\n
o Comparative benchmarking across models and configurations
\n
• Define metrics for:
\n
o Accuracy, relevance, and consistency
\n
o Cost efficiency and latency
\n
o User trust and explainability
\n
• Partner with engineering and risk teams to ensure:
\n
o Observability into AI behavior
\n
o Safe deployment and controlled experimentation
\n
o Continuous improvement loops based on real usage
\n
Platform Enablement & Collaboration
\n
• Work closely with:
\n
o Platform engineering teams
\n
o Product and design partners
\n
o Consulting and delivery leaders
\n
• Provide technical guidance on:
\n
o How AI capabilities should be embedded into platform features
\n
o Where AI adds leverage vs. complexity
\n
• Support enablement through:
\n
o Technical documentation and reference architectures
\n
o Internal education and design reviews
\n
o Advisory support for high-impact use cases
\n
Governance & Responsible AI
\n
• Define technical guardrails that support:
\n
o Security, privacy, and data residency
\n
o Responsible AI principles
\n
o Regulatory and client requirements
\n
• Ensure AI systems are:
\n
o Explainable where required
\n
o Observable and auditable
\n
o Designed for controlled evolution over time
\n
What Success Looks Like
\n
• The platform consistently adopts relevant AI innovations without destabilizing
\n
delivery.
\n
• AI behavior is predictable, testable, and trusted by consultants and leadership.
\n
• New models, agents, and techniques can be introduced rapidly through well
\n
defined abstractions.
\n
• AI capabilities directly improve delivery quality, speed, and consistency across
\n
engagements