At TKXEL, I work on CanvsAI, an enterprise customer insights platform that turns qualitative customer feedback into actionable business intelligence. It serves large enterprises including Netflix and Udemy.
I built production Agentic AI workflows with LangGraph’s Orchestrator-Worker pattern and RAG pipelines using Pinecone. I also use LangSmith for tracing, prompt management, evaluations, and human feedback.
At DATICS AI Pvt. Ltd, I developed full-stack applications and worked across monolithic and microservices architectures. I used Python backend frameworks and asynchronous processing with Celery and RabbitMQ.
On Secure Tax, I restructured a Django monolith into microservices for a US property tax collection application, using Celery and RabbitMQ for background tasks. I’ve also contributed to both Python APIs and React interfaces, integrating OpenAI GPT and Claude models.
