Responsibilities
- Collaborate closely with internal departments to observe workflows, identify inefficiencies, and pinpoint areas where AI can deliver significant improvements.
- Serve as the frontline AI specialist by joining team meetings, process evaluations, and planning discussions to uncover new AI use cases.
- Convert business and operational needs into clear technical requirements for AI systems, facilitating communication between engineers and non-technical stakeholders.
- Manage full-cycle deployment of AI tools within operational settings, from initial scoping and design to development, testing, rollout, and ongoing refinement.
- Offer direct technical support for live AI models, diagnosing issues and ensuring stable, high-performing operation within internal environments.
- Lead user adoption efforts by training staff on AI tools, updated workflows, and recommended practices for effective use.
- Record real-world AI implementations, including patterns, challenges, and outcomes, to guide future product development and broader team rollouts.
- Develop and maintain enterprise-grade features powered by large language models, such as natural language querying and AI-guided processes.
- Construct agent workflows that enable multi-stage reasoning, tool invocation, error recovery, result summarization, and final response creation.
- Establish and manage standardized tool definitions for LLM integrations, specifying names, inputs, outputs, descriptions, and safety constraints.
- Build logic that routes LLM tool requests to backend functions, executes them, manages errors, and returns processed results to the conversation flow.
- Enable transparent AI interactions by allowing users to review executed steps, generated queries, data outputs, chart suggestions, and final explanations.
- Craft specialized prompts and instruction templates tailored to managed services and pharmaceutical data analysis workflows.
- Refine prompt designs to ensure consistent, concise, and predictable outputs across different interaction modes like quick queries, agent-based tasks, and chat.
- Apply knowledge of LLM limitations including context length, token usage, conversation memory, and security risks such as prompt injection.
- Evaluate and enhance model performance using test suites, failure diagnostics, and iterative adjustments to prompts and logic.
- Implement strategies to reduce token consumption and operational costs through efficient result formatting, context management, and model selection.
- Develop robust chat interfaces that handle user input, support streaming or step-by-step responses, maintain history, and display tool outputs clearly.
- Work with product and operations teams to ensure AI outputs are clear, reliable, and useful for both technical and business audiences.
- Incorporate visual recommendations, tabular results, SQL previews, and execution logs into the AI interface for better transparency.
- Design error handling mechanisms for common failure points such as timeouts, invalid tool calls, malformed JSON, failed queries, and incomplete agent outputs.
- Coordinate with site reliability and security teams to establish secure, observable, and maintainable deployment practices for AI systems.
- Architect and support deployment strategies in AWS environments, including containerized applications, API-driven services, and secure access to identity and data platforms.
- Develop backend components and services using Python and modern API design principles.
- Follow software engineering best practices including version control, code reviews, automated testing, and documentation for reliable releases.
Work Arrangement
Hybrid
