AI Developer

Job description

Key Responsibilities Workflow Digital Twin Development ? Analyze and deconstruct operations user workflows (tasks, decisions, exceptions, dependencies) ? Build digital representations of workflows (decision trees, state machines, agent flows) ? Translate business processes into automated, executable logic ? Capture and model edge cases, branching logic, and exception handling Agentic AI & Automation ? Develop AI agents/copilots that mimic operations user behavior and decision-making ? Implement orchestration using tools such as: • Copilot Studio / Azure AI Foundry • LangChain / Semantic Kernel (or similar frameworks) ? Enable human-like task execution (data retrieval, validation, enrichment, updates) AI/ML & Decisioning ? Build models for: • Classification and routing • Anomaly detection • Recommendation/decision support ? Integrate LLMs for reasoning, summarization, and workflow execution assistance Data & Systems Integration ? Integrate with enterprise systems (APIs, databases, file systems, M365, CRM) ? Build pipelines for ingesting structured and unstructured operational data ? Ensure data consistency and traceability across workflow steps Testing & Validation ? Validate Digital Twin accuracy against actual user behavior and outcomes ? Perform side-by-side comparisons (human vs digital twin output) ? Continuously refine logic to improve fidelity and performance Monitoring & Optimization ? Track workflow execution metrics (accuracy, cycle time, exceptions) ? Improve automation coverage and reduce manual intervention ? Support production deployments and troubleshoot issues Required Qualifications ? 3?5 years? experience in software development, automation, or AI/ML ? Strong Python skills and experience building APIs and automation workflows ? Experience with: • Process automation (Power Automate, BPM tools, or custom pipelines) • REST APIs and system integration • Data processing (Pandas, SQL) ? Ability to translate business workflows into technical logic Key Skills ? Strong logical and structured thinking (if/then, branching, dependencies) ? Ability to mimic human decision-making in code ? Debugging complex workflows and edge cases ? Collaboration with non-technical stakeholders Success Metrics ? % of ops workflows successfully replicated and automated ? Reduction in manual effort and cycle time ? Accuracy of digital twin vs human execution ? Reduction in exceptions and rework Example Work ? Build an AI agent that replicates trade reconciliation workflows ? Automate document review and data extraction + validation steps ? Mimic email-driven ops processes (triage ? action ? update systems) ? Create decision engines for exception handling and routing