Data & AI tech Manager/Data & AI Tech Manager

Job description

Position Principal Responsibilities

Data Management and Extraction

• Maintain an inventory of data assets, including databases, data sources, and formats, ensuring that all data is easily accessible and well-organized.
• Design and implement processes for data extraction from various internal and external sources, ensuring accurate and timely retrieval of information.

Data Integration

• Collaborate with other functions to design and implement system integration solutions that ensure seamless data flow between various applications and platforms.
• Manage and optimize databases to ensure efficient data storage, retrieval, and processing, supporting the organization’s data needs.

Data Analytics and Tools

• Collaborate with business units to gather data requirements, providing support for data analysis via reporting tools, system integration, and AI applications.
• Work closely with various departments to understand their data needs and provide support for data-related projects and initiatives.
• Oversee the selection, implementation, and management of data analytical tools and platforms (e.g., BI tools, data visualization software) to enhance data analysis capabilities across the organization. Agentic Architecture & System Design

Design Multi-Agent Systems

• Architect scalable multi-agent orchestration frameworks where specialized agents collaborate to solve complex, multi-step business problems.

Build Advanced RAG Pipelines

• Lead the development of next-gen Retrieval-Augmented Generation (RAG) systems that go beyond simple vector search, integrating Knowledge Graphs, hybrid search, and re-ranking models to provide agents with precise, context-aware grounding.

Memory & State Management

• Design sophisticated long-term and short-term memory mechanisms (using vector DBs, SQL, or key-value stores) to allow agents to maintain context across long-running sessions and learn from past interactions. Application Engineering & Product Delivery

End-to-End Agent Lifecycle

• Own the full software development lifecycle (SDLC) for AI applications, from prototype to production, ensuring agents are not just demos but reliable, latency-optimized products.

Human-in-the-Loop (HITL) Interfaces

• Build intuitive UI/UX patterns for human-agent collaboration, including approval gates, intervention points, and feedback loops humans guide or correct agent actions before final execution.

Evaluation & Testing Frameworks

• Establish rigorous automated testing suits for agents to measure success rates, task completion accuracy, hallucination frequency, and tool invocation reliability.

Performance Optimization

• Optimize application performance by implementing caching strategies, prompt compression, model routing, and asynchronous processing to reduce latency and token costs. Reliability, Safety & Governance (Agent-Specific)

Guardrails & Safety Layers

• Implement advanced guardrail systems to prevent agents from taking unauthorized actions, accessing sensitive data, or entering infinite loops. This includes input/output filtering and constraint enforcement.

Deterministic Workflow Enforcement

• Balance probabilistic LLM reasoning with deterministic workflow engines to ensure critical business processes remain predictable and auditable.

Observability & Debugging

• Deploy comprehensive observability stacks specifically for agentic flows, tracing decision paths, tool calls, and reasoning steps to rapidly debug failures in complex autonomous chains.

Cost & Token Governance

• Monitor and optimize token consumption and compute costs associated with high-frequency agent operations, implementing budget limits and efficiency protocols. Technical Leadership & Innovation

Framework Selection & Strategy

• Evaluate and select the best open-source and proprietary agent frameworks and define the team’s technical stack.

POC to Production Pipeline

• Rapidly prototype new agent concepts and define the clear criteria and engineering standards required to graduate them into mission-critical production applications.

Cross-Functional Integration

• Collaborate closely with product managers and domain experts to translate vague business needs into structured agent specifications and actionable user stories.

Team Upskilling

• Mentor engineers on prompt engineering patterns, chain-of-thought reasoning, few-shot learning, and the nuances of building non-deterministic software systems. ------------------------------------------- Principal Responsibilities Data Management and Extraction

Team Upskilling

• Mentor engineers on prompt engineering patterns, chain-of-thought reasoning, few-shot learning, and the nuances of building non-deterministic software systems.