Yash Kumar

Designed and implemented AI-driven financial prediction models using Python and TensorFlow, enabling more accurate risk assessments and portfolio optimizations. Built scalable AI microservices with Flask and FastAPI for integrating ML models into enterprise systems. Developed NLP pipelines to extract and analyze regulatory documents, improving compliance monitoring efficiency. Collaborated remotely with data scientists and engineering teams in agile sprints to deliver iterative AI feature enhancements aligned with evolving business needs. Created comprehensive unit and integration testing frameworks to validate model accuracy and system stability, reducing production bugs and downtime.

Designed and developed a complete SaaS platform integrating agentic AI agents to automate sales lead generation and CRM workflows, significantly improving client engagement and operational efficiency. Engineered backend services and APIs supporting AI prompting and dynamic user interaction with daily task reminder agents, enhancing productivity across internal user groups and customer-facing portals. Led architecture and implementation of cloud deployment pipelines using container orchestration and serverless functions, reducing operational costs by 30% while ensuring high availability. Managed robust databases schemas for multi-tenant SaaS applications optimized for concurrency and rapid data access. Automated workflow processes including message-based WhatsApp notifications using Twilio APIs, increasing automation outreach by 40%. Collaborated with AI research teams to integrate natural language processing-based AI prompting capabilities, improving predictive response accuracy and pe

Engineered ML classification models using Python and scikit-learn to automate supplier categorization, reducing manual review times by 40%. Implemented ETL pipelines to preprocess large-scale retail datasets for AI analysis, improving data quality and consistency supporting demand and forecasting tasks. Developed NLP-enabled sentiment analysis tools for customer feedback, enabling real-time strategy adjustments. Built REST APIs to expose AI model inference capabilities to internal applications, enabling seamless integration between data science and software engineering teams. Automated hyperparameter tuning workflows to improve classifier accuracy with minimal human intervention. Conducted performance profiling and optimization on AI pipelines, accelerating model training and inference phases.