Projects Portfolio

Technologies Used:

Key Highlights:

• Built with Streamlit, Langgraph, and OpenAI, leveraging GPT-4o-mini for natural language processing and Tavily API for real-time web searches
• Integrated vector databases for efficient storage and retrieval of knowledge, enabling intelligent decision-making
• Designed intuitive chat interface with session memory, highlighted keywords, and clear formatting
• Enabled users to upload PDFs and receive simplified answers, improving learning efficiency by 30%

Technologies Used:

Key Highlights:

• Built a fully automated sales agent system using Streamlit and LangGraph with GPT-4o-mini for intelligent analysis
• Integrated multiple department-specific Excel workbooks with SharePoint for seamless real-time updates
• Enabled dynamic query responses, report generation, and data visualization for faster decision-making
• Applied 4-tier analytics (Descriptive, Diagnostic, Predictive, Prescriptive) to provide actionable insights from sales KPIs and trends.

Technologies Used:

Key Highlights:

• Enabled quick and accurate insights by automating data extraction and summarization, reducing manual effort by 40% and improving processing speed by 50%
• Implemented advanced NLP techniques to handle complex document layouts, achieving 95% accuracy in data extraction
• Integrated the solution into existing workflows, reducing turnaround time from hours to minutes
• Built user-friendly interface increasing adoption by cross-functional teams by 30%

Technologies Used:

Key Highlights:

• Built with Python (Flask) and JavaScript, leveraging Web Speech API for speech-to-text and deep_translator for real-time translation
• Implemented text-to-speech functionality using gTTS for audio playback of translated text
• Designed mobile-friendly, responsive interface using Tailwind CSS
• Supported 20+ languages including English, Spanish, French, Hindi, Arabic, Chinese, and Japanese
• Reduced communication barriers in healthcare settings

Technologies Used:

Key Highlights:

• Integrated Large Language Model to automatically summarize reports
• Made data analysis faster and more accessible for non-technical users
• Streamlined querying and reporting process, reducing time to generate insights by 40%
• Designed intuitive interface improving user adoption by 25%
• Enhanced data accessibility and decision-making efficiency

Technologies Used:

Key Highlights:

• Designed and trained neural network model for efficient facial feature extraction and recognition
• Optimized performance for diverse lighting and environmental conditions
• Integrated solution with live camera feed and database for face storage and retrieval
• Achieved high accuracy in face detection and identification for law enforcement use
• Leveraged Python and AI frameworks to ensure scalability and seamless deployment
• 📰 Featured in national media coverage for innovative AI implementation

Technologies Used:

Key Highlights:

• Designed and implemented a centralized data warehouse integrating multiple transactional databases
• Developed ETL pipelines to clean, transform, and load data efficiently
• Created analytical reports and dashboards for business insights
• Optimized query performance and improved reporting speed by 35%
• Enabled strategic decision-making through consolidated data analysis

Technologies Used:

Key Highlights:

• Built a multi-page interactive Power BI dashboard with real-time data visualization
• Performed in-depth analysis of sales, customers, and products for 2016-2017
• Integrated SQL Server data with Power BI for automated reporting
• Enabled dynamic slicing/dicing of data by year, month, product, and customer attributes
• Supported business strategy with actionable insights and KPI tracking

Technologies Used:

Key Highlights:

• Supports multi-modal communication: text input, voice notes, and voice responses. :contentReference[oaicite:1]{index=1}
• Maintains conversation state over WhatsApp with a relational database backend (PostgreSQL) for persistence. :contentReference[oaicite:2]{index=2}
• Implements transcription of voice messages (e.g., via Whisper or equivalent) and synthesised voice responses. :contentReference[oaicite:3]{index=3}
• Leverages LangGraph’s agent-orchestration to manage multi-step flows and memory. :contentReference[oaicite:4]{index=4}
• Demonstrates WhatsApp use-case for conversational AI beyond simple chatbots—designed for empathy and long-term context.

Technologies Used:

Key Highlights:

• Enables natural-language querying of structured data using RAG pipelines.
• Supports retrieval of relevant database context and generation of responses grounded in the data.
• Bridges LLMs with actual database back-ends (SQL) rather than purely document retrieval.

Technologies Used:

Key Highlights:

• Facilitates scalable resume intake and management for HR operations.
• Implements parsing and indexing of resumes for candidate profiling.
• Potentially integrates screening, sorting, and matching capabilities.

Technologies Used:

Key Highlights:

• Automates PDF → structured database insertion, eliminating manual data entry.
• Uses AI/NLP for extracting fields from unstructured PDF content.
• Enables downstream analytics or BI reporting via structured data store.