Vikesh Patel

About

ML/NLP practitioner who has delivered production systems in regulated domains, including a healthcare compliance pipeline using RAG (GPT-4/Claude) plus TF-IDF retrieval that increased document review throughput 4.5x. Also has hands-on experience improving fraud detection data quality via entity resolution (Levenshtein, Dedupe.py) validated with A/B testing, and building scalable, monitored workflows with Airflow, CI/CD, and AWS SageMaker.

Experience

Education

Key Strengths

• Built RAG-based compliance document processing using GPT-4/Claude with TF-IDF retrieval
• Improved healthcare document review throughput by 4.5x
• Entity resolution for financial fraud datasets using fuzzy matching (Levenshtein) and Dedupe.py
• Validated data cleaning/entity resolution via A/B tests against manually curated ground truth; reduced false positives and improved model accuracy
• Implemented semantic search with sentence embeddings + Redis vector store; improved retrieval accuracy via domain fine-tuning
• Production-grade Python data workflows with Airflow, monitoring (Prometheus/Grafana), and CI/CD (GitHub Actions)
• Managed production changes with A/B-tested SageMaker migration enabling zero-downtime rollout

Reference Highlights

• strong expertise in large-scale ML/NLP solutions
• builds scalable AI systems (fraud detection, chatbots, ML platforms)
• effective at addressing model drift and latency
• delivers robust, reliable data pipelines
• automates feature engineering
• significantly reduces manual work (80 hours/month)
• improves model deployment efficiency
• clear communicator of complex technical concepts
• collaborates seamlessly with cross-functional teams
• leads AI model governance discussions
• aligns technical strategy with compliance requirements
• strong ownership and accountability
• leads from the front in decision-making
• delivers production-ready solutions
• drives measurable delivery improvements (60% reduction in deployment time)
• strong leadership
• highly valued; would be rehired without hesitation