Portfolio
What Service You Will Get From Me
A showcase of innovative solutions built with cutting-edge AI, NLP, and search technologies.
AI-Powered Global Search System (ServiceNow)
Duration – 3 Years
Migrated a legacy Apache Solr-based system to AI Search within ServiceNow, significantly enhancing performance, scalability, and manageability. Developed robust RESTful APIs to enable real-time data ingestion across global ServiceNow portals. Integrated Azure OpenAI to deliver intelligent, generative summaries of search results. Implemented a caching layer to optimize LLM response times and improve user experience.
Tech Stack: Python, Java, AI Search, Azure OpenAI, ServiceNow
Job & Resume Search Engine
Duration – 1 Year
Designed and deployed a scalable job and resume search system using Elasticsearch hosted on AWS. Implemented advanced Boolean and predictive search logic along with multilingual query support to cater to a global user base. Enhanced search quality using language-specific analyzers and optimized the system for performance under high data volume.
Tech Stack: Elasticsearch, Python, Java, AWS, REST APIs
Blockchain-Based Reward System (OI Coin)
Duration – 6 months
Implemented a secure blockchain-based rewards ecosystem using Binance Smart Chain and Solidity. Developed smart contracts to allocate OI Coins to users based on activity. Integrated real-time transaction monitoring and ensured secure handling of user data and crypto assets through REST APIs.
Tech Stack: Solidity, Binance Smart Chain, Python, AWS
Multilingual Chatbot for 50+ Websites
Duration – 3 Years
Developed and deployed an NLP-driven chatbot platform across 50+ websites, supporting 8+ languages. Enabled dynamic entity, intent, and slot recognition using SpaCy and Scikit-learn. Integrated semantic features like spell correction and sentence auto-completion. Managed scalable deployment using Kubernetes and GCP.
Tech Stack: Python, Scikit-learn, SpaCy, Kubernetes, GCP
LLM-Powered Summarization for Auto Fixes
Duration – 3 Years
Used Llama2 models to automate descriptive summarization of automotive repair logs. Extracted key insights such as replaced parts and service recommendations. The output was pre-processed and indexed into Solr for instant retrieval, with continuous model tuning for quality optimization.
Tech Stack: LLMs, Python, Solr, MySQL, Redis
Vehicle Trouble Code Search Engine
Duration – 3 Years
Built a powerful Solr-based search solution for 50M+ vehicle trouble codes. Introduced clustering to organize similar codes and NLP-based entity extraction. Added features like autocorrect and search suggestions, and developed a recommendation engine based on user history.
Tech Stack: Solr, Python, AWS Beanstalk
Unified Search for Communication Suite
Duration – 4 Years
Engineered a unified search engine across integrated communication tools including email (Zimbra), chat (Prodigy Server), and project boards (Redmine). Created a custom Solr integration to handle indexing and querying, with both real-time and bulk ingestion pipelines using Redis Pub/Sub.
Tech Stack: Solr, Java, Spacy, Redis, Zimbra, Redmine
Knowledge Extraction System Using NLP
Duration – 4 Years
Developed a comprehensive NLP platform that combined machine learning and rule-based algorithms. Supported features like summarization, sentiment analysis, and keyword extraction. Enabled dynamic entity recognition and SPARQL-based querying over structured knowledge graphs.
Tech Stack: Python, Java, SPARQL, Apache Solr, Sesame Server
Search & Recommendation Engine for Recruitment
Duration – 1 Years
Designed Solr-based indexing and search APIs for job listings and resumes. Leveraged NLP to enhance candidate-job matching, and implemented Firebase for real-time job notifications. Included entity recognition and recommendation logic for improved relevance.
Tech Stack: Python, Java, Solr, Firebase, NLP