Hi, I'm Prasanth.

Current work centers on organizing product intelligence workflows and validating the quality of outputs used for discovery and recommendations.

• Contribute to product data organization, product capture, matching, search, ranking and recommendation workflows.
• Analyze product attributes, reviews and merchant information for product discovery and consumer-focused insights.
• Support product benchmarking and comparative analysis across similarities, rankings, recommendation signals and data quality.
• Analyze product-preference and buying-habit signals for consumer insights and personalized recommendations.
• Evaluate search, ranking, matching and recommendation outputs and identify improvement opportunities.
• Validate product records to improve accuracy, consistency and reliability.
• Map business/technical workflows, requirements, dependencies and integration points.
• Perform API testing and collaborate using AI/ML, APIs, MongoDB and related technologies.
• Track technical tasks, testing activities, action items and follow-ups across ongoing development.

Worked on production Generative AI and NLP systems for enterprise knowledge management and intelligent document processing.

• Architected and deployed Generative AI applications using GPT-4, LLaMA 2 and Mistral.
• Designed RAG pipelines with LangChain, LangGraph, Pinecone and FAISS, plus multi-agent workflows for task planning, tool invocation and automation.
• Engineered NLP pipelines for document classification, sentiment analysis, NER and summarization.
• Implemented ingestion, chunking, metadata enrichment, embeddings and vector indexing across 50,000+ enterprise documents.
• Improved classification F1-score from 0.74 to 0.86 through feature engineering, tuning, cross-validation and A/B testing.
• Optimized inference using prompt optimization, caching, quantization and efficient model serving to reduce latency and infrastructure costs.
• Collaborated with product managers, data scientists and software engineers to deploy AI microservices on AWS.
• Performed testing, debugging, performance optimization and model evaluation for production reliability.
• Translated business requirements and AI use cases into scalable technical solutions.

Supported enterprise workflows by integrating NLP models and automating data-quality and operational processes.

• Integrated NLP models for automated ticket classification and priority routing.
• Reduced manual triage time by 25% across 10,000+ monthly tickets.
• Developed Python preprocessing for structured and unstructured datasets.
• Built anomaly-detection pipelines using Python and scikit-learn.
• Designed automated validation workflows and optimized SQL/ETL workflows.
• Performed EDA to identify trends and anomalies for business stakeholders.
• Automated repetitive tasks using Python, saving over 2,000 hours of manual effort annually.
• Collaborated cross-functionally, participated in Agile/Scrum and documented technical designs.