KULDEEP

Currently working on Generative AI initiatives as part of the Velocity AI project. Key responsibilities include:

Implemented Retrieval-Augmented Generation (RAG) pipelines to improve contextual response quality using internal and external data sources.

Developed backend services using FastAPI, enabling scalable and efficient API integration across AI modules.

Integrated with multiple LLM platforms, including Ollama, Azure OpenAI, Amazon Bedrock, and Vertex AI, to support diverse generative AI use cases.

Connected with Jira APIs to automate the creation of Epics and user stories, enhancing Agile planning workflows.

Integrated ChromaDB to store and manage vector embeddings, supporting semantic search and document retrieval.

• Company Overview: A company specializing in medical equipment and healthcare solutions.
• Digital Microscope Whole Slide Imaging: Developed REST APIs with Django for automated slide scanning and analysis, enabling pathologists access to imaging data.
• Blood Cell Classification: Developed an automated blood cell classification system using YOLOv8 for real-time detection and classification of blood cell types. Integrated advanced image processing techniques, including histogram equalization, CLAHE, Gaussian blur, and morphological operations, to improve image clarity and reduce noise. Leveraged YOLOv8's instance segmentation for precise cell localization, isolating individual WBC, RBC, and platelets, while semantic segmentation mapped broader cell types across the sample. This approach significantly enhanced accuracy in blood cell subtype categorization, essential for efficient medical diagnostics.
• Autofocus Algorithm: Built a dynamic autofocus algorithm utilizing image sharpness metrics, ensuring optimal image clarity across specimen types.
• Slide Scanning & Stitching: Created an algorithm for whole smear slide scanning, stitching high-resolution images for a comprehensive digital representation.
• A company specializing in medical equipment and healthcare solutions.

• Company Overview: A technology solutions provider.
• Increased IT support efficiency by developing a Rasa-based chatbot integrated with WhatsApp and Teams.
• Implemented multiple chatbots to improve customer interaction efficiency, leveraging APIs and ML for seamless integration and robust response management.
• A technology solutions provider.

• Company Overview: A company focused on IoT and Predictive maintenance technology.
• Machine Fault Detection and Prediction: Developed and implemented algorithm to detect and categorize faults in industrial machinery.
• Analyzed vibration data, temperature, and other sensor metrics to predict potential failures, helping reduce downtime and maintenance costs.
• Threshold and Baseline Alert System: Designed threshold-based alert mechanisms that notify operators of potential issues based on abnormal sensor readings, with baseline adjustments to accommodate equipment-specific variations.
• Router Website Development: Created a web-based interface using Flask, SQLite, and Python, providing real-time monitoring and network management of connected sensors. Enabled easy access to sensor data and network configurations for proactive maintenance.
• A company focused on IoT and Predictive maintenance technology.