Enterprise AI search which connects to various source systems and internal knowledge base to provide search with better accuracy and relevancy.

• Built new indexing and search system from scratch which connects to various source systems including newly adopted Kontent.ai CMS and stores vectorized documents in Elasticsearch.
• Used SBERT and advanced NLP models for text embedding, text classification, spell check and other NLP tasks.
• Optimized code of indexing and searching using concurrency and multi-threading concepts which reduced the latency of the AI search API by 30-40% and indexing jobs by 50%.
• Increased the overall search relevancy and accuracy by 40% by following recommended strategies and fine tuning data which helped the users find accurate information in search results.

Analytics dashboard for chatbot to display various statistical data and KPI metrics with visualizations.

• Being a solo developer in the team, created the dashboard for chatbot from scratch using Vue.js and Express.js.
• Written several optimized MongoDB aggregations for fetching data on millions of records for better performance.
• Created Jupyter notebooks using Pandas & Numpy for better analysis and measuring the efficiency of chatbot.

Chatbot developed for intranet portals built on IBM Watson which uses NLU to provide better experience for associates.

• Responsible in full stack development of the application using Angular, Socket.io and Node.js.
• Created several RESTful API s which talks to multiple backends and integrated with many enterprise applications like Peoplesoft, Ariba, Workday, Skillsoft, etc.