Debanka Das

Lead architecture and technical strategy for enterprise agentic AI solutions across Workday HCM and Financial Management environments, designing multi-agent architectures and defining governance and security boundaries.

Built an explainable model-routing platform that selects models per request based on task characteristics, capability constraints, quality, cost, latency and provider conditions, with benchmark and adaptive-learning infrastructure.

Built a self-learning, security-first open-source AI agent OS with persistent memory, reflection and evolution loops, model routing, tool execution and autonomous workflows.

Built DynaTune, a no-code fine-tuning platform, and a production-grade enterprise search and conversational AI platform indexing approximately 3.4 million documents. Designed intent-aware retrieval and a time-series forecasting platform.

Owned technical direction for enterprise AI-driven BSS initiatives, architected an OpenAI-powered RAG solution reducing average handling time by 35%, and led model selection across LLMs. Provided architecture leadership on Kubernetes and CI/CD.

Led CRM and billing transformations for a tier-1 telecom operator, designing Kafka and Airflow-based data pipelines that reduced order processing time by 30% and improved billing accuracy.

Owned solution design for large-scale CRM, ordering and billing platforms serving millions of subscribers, and developed technology roadmaps across CRM, OMS, charging, invoicing and revenue-management integrations.

Progressed from software development into business analysis and solution ownership across enterprise BSS programs, leading ordering and billing enhancements that reduced order fallout by 15% and improved revenue assurance.

Bachelor of Computer Applications (BCA) from Techno India, completed in 2010.