This is the second paragraph of your amazing article.
(adsbygoogle = window.adsbygoogle || []).push({});
This is the third paragraph where the content continues.
This is the second paragraph of your amazing article.
(adsbygoogle = window.adsbygoogle || []).push({});
This is the third paragraph where the content continues.
Built an AI-driven driver verification & onboarding system
automating document validation and identity checks using
structured parsing and confidence-based routing, reducing manual review by ~75% and improving onboarding speed. Developed a hybrid support assistant using Rasa for intent classification and rule-based flows, with LLM routing for complex queries. Integrated with backend APIs via FastAPI and
implemented confidence-based fallback handling.
Improved domain-specific LLM reliability through prompt
iteration, structured test cases, and response validation guardrails, increasing consistency and reducing incorrect outputs.
Implemented a multi-document RAG pipeline ingesting PDFs and internal knowledge content, using embedding-based
retrieval with metadata filtering to improve answer relevance
and reduce manual search effort. Refactored backend services into modular components with
schema validation, centralized logging, and retry mechanisms, reducing deployment issues and improving maintainability.
Restricted contact info. You need company account to have access.
