Manobhav Joshi

Results-oriented Technology Analyst with over 3.9 years of experience in developing and deploying AI-driven solutions. Specializes in generative AI, computer vision, and large language models (LLMs). Solid grasp of data munging techniques. Demonstrates expertise in designing, building, and optimizing high-performance machine learning models on cloud platforms such as Azure and AWS. Experienced in creating scalable generative AI systems that effectively address real-world business challenges. A strong communicator, fast learner and team player who excels in collaborating and mentoring with cross-functional teams to deliver innovative AI solutions that drive impactful results while also considering risk management.

• Generative AI Document Generator:
  Developed a Retrieval-Augmented Generation (RAG) based document generation system that streamlined the creation of critical business documents, improving client efficiency by reducing document preparation time by over 40%.
  Technologies Used: Python, Azure OpenAI, LangChain, ReactJS, FastAPI.

• AI-Powered Document Summarizer and PPT Generator:
  Built AI-powered summarization and automated presentation generation tools utilizing generative AI models to summarize complex data, reducing processing time by 40% and enhancing client presentations.

• AI Agent Integration (Conversational AI & Database Query System with RAG):
  Developed advanced conversational AI agents using Retrieval-Augmented Generation (RAG), capable of interpreting user queries, generating dynamic SQL queries, and retrieving data from databases in real-time.
  Integrated Natural Language Processing (NLP) with SQL databases and vector databases to provide context-aware responses, significantly enhancing client data accessibility and accuracy.
  Improved decision-making efficiency by providing instant, precise data insights through intuitive chatbot interfaces.
  Technologies Used: Python, LangChain, Azure OpenAI, MSSQL, Elasticsearch, REST APIs, Docker.

• AI-Driven Data Pipelines:
  Developed and deployed scalable AI-driven data pipelines to ensure robust data processing workflows, enhancing the stability and accuracy of machine learning models in production.

• Supply Chain and Warranty Claims Chatbot (RAG-Based):
  Currently developing a conversational chatbot leveraging Retrieval-Augmented Generation (RAG) for supply chain and warranty claim processes, significantly streamlining customer interactions and improving operational efficiency by approximately 35%.
  The chatbot effectively handles complex queries, retrieves accurate warranty and supply chain data instantly, and enhances decision-making processes.
  Technologies Used: FastAPI, Azure OpenAI, Elasticsearch, Python, MSSQL, Data lakes, ReactJS, Docker.

• AI-Powered Analytical Dashboards:
  Designed interactive, user-centric dashboards with ReactJS frontend and PostgreSQL backend to deliver real-time analytics.
  Improved data visualization capabilities, enabling stakeholders to gain actionable insights and enhancing overall decision-making quality.

• AI Feature Development and Deployment:
  Developed end-to-end AI-powered features from data pipelines to frontend integration, improving system efficiency and user experience.

• Ensure Data Encryption & Protection – Use secure encryption techniques to protect sensitive data.
• Access Control & Authentication – Implement robust role-based access controls to prevent unauthorized data usage.
• Diverse Training Data – Use balanced datasets to prevent bias in AI predictions.
• Continuous Model Monitoring – Track AI model performance over time and retrain when necessary.
• Automated Retraining Pipelines – Use MLOps for automated model updates.
• A/B Testing & Validation – Validate models before full-scale deployment.
• Resilient AI Workflows – Implement failover mechanisms to ensure AI system availability.
• Optimized Cost Management – Monitor AI compute resource usage to prevent excessive operational costs.
• Energy-Efficient AI Models – Optimize AI models that require less processing power so to consume less power.