Dipendra Shekhawat

AI & Automation Leader | AI Platform Architecture & Governance | Agentic AI Delivery

Profile summary

Career highlights

Key skills

Professional experience

Steered the GenAI and AI digital automation strategy while managing a team of 15 to 17 engineers. Defined governance frameworks for low-code and pro-code AI platforms through close partnership with business leaders to deliver production-grade enterprise solutions.

• •Architected and delivered production-grade GenAI and agentic AI programs including the Validation Agent and Agentic Twin.
• •Governed the accelerator platform framework by mapping approved platforms to complexity tiers and user personas across a portfolio.
• •Integrated production controls into GenAI solutions such as human-in-the-loop validation, safety screening, and role-based access.
• •Conducted technical feasibility assessments and architecture reviews for enterprise intelligent automation initiatives.
• •Mentored automation and AI engineers across solution design, engineering quality, and professional development.

Shaped solution delivery strategy and directed engineering operations for a team of 8 to 10 professionals spanning discovery, requirements analysis, design, development, testing, deployment, and go-live support. Directed enterprise architecture standards on the Automation Anywhere platform, collaborating directly with customers and internal stakeholders to resolve complex implementation issues and communicate technical recommendations.

• •Architected enterprise solutions on Automation Anywhere platforms and components including A360, AARI, IQ Bot, Bot Insight, C#, SQL Server, JavaScript, Python, and conversational AI.
• •Led migrations of automation workloads from v11.x to A360, applying platform best practices and reusable design patterns.
• •Performed architecture reviews, code reviews, troubleshooting, and technical mentoring for developers and project teams.

Directed technical design and product roadmaps across engineering teams for digital workplace products and intranet solutions, owning architectural decisions and database logic.

• •Delivered full-stack web applications using ASP.NET, C#, SQL Server, JavaScript, ReactJS, Redux, ES6, and TypeScript.
• •Built Serverless components in C# on AWS including Lambda and S3 as part of application delivery.
• •Built an automated solution using UiPath for internal teams.

Directed technical architecture for enterprise applications, coordinating software delivery across client stakeholders and internal engineering teams while guiding junior staff.

• •Built and maintained enterprise applications using Classic ASP, ASP.NET, C#, SQL Server, MVC, WCF, HTML, and CSS.
• •Implemented database logic and supported data migration activities using SQL Server and DTS.
• •Delivered solution demonstrations to clients and supported knowledge transfer and mentoring for junior engineers.

• •Built an internal web application to support team operations.
• •Processed geographic data inputs to update digital mapping databases according to project specifications.

Projects

Architected and built an end-to-end GenAI automation platform to eliminate the manual, ~30,000-document/year validation cycle required for new equipment introduction (previously performed by 1,000+ people across sites). Designed a layered architecture spanning knowledge capture (RAG over Azure AI Search/Glean), curation and persistence (Azure Storage, Azure ML), agent orchestration, and a serving layer (FastAPI, Kong Gateway) in front of a UI and an Automation Anywhere execution layer. Initially built orchestration with Azure Prompt Flow, then re-architected it onto LangChain and LangGraph for more flexible, stateful agentic control flow. Integrated Azure AI services (Document Intelligence, Content Safety, Translator) for document generation and scoring, with Azure OpenAI and Azure ML models served through a centralized gateway.

Architected the platform's shared foundation on Microsoft Copilot Studio and M365 Agent Builder: "Twinny," a routing/orchestration agent that acts as the program's single entry point, interprets user intent, and directs requests to the correct specialist agent, running on common Power Platform infrastructure (Power Automate, Dataverse) shared across all agents. Built four specialist agents on top of this shared platform: A Problem-Solving Twin (DMAIC documentation and deviation management), a Meeting Twin (pre-reads and post-meeting notes/action items for governance meetings), and a Mentor Twin (manager 360 reviews and reportee performance assessment) - demonstrating how a shared orchestration and infrastructure layer accelerates delivery of individual AI applications.

Architected a POC-stage agentic solution to compress the Contract Manufacturing Organization (CMO) on boarding process, which spans 211 tasks across 9 departments and typically takes 16-17 weeks for lower-complexity scenarios and up to 26 weeks for complex ones (16 distinct on boarding scenarios in total), driven largely by manual data and document collection, verification, cross-functional process alignment, and task coordination. Architecture: exposed SAP ERP on boarding data to Databricks as the central data layer; built n8n workflows to fetch and orchestrate data across the majority of workflows (covering the 16 on boarding scenarios); designed a front-end UI layer embedded in Microsoft Teams to collect data and support human-in-the-loop approvals and interactions at key decision points. Built as a reusable pattern intended to scale beyond CMO on boarding to other similarly structured, multi-department on boarding/readiness processes.

Education

72%

Certifications