• Designed, developed, and shipped production-grade voice AI agents for appointment scheduling, clinical documentation, and insurance verification, deployed on GCP (Cloud Run/GKE) with autoscaling and Cloud Logging/Monitoring, handling 10K+ calls per week.

• Researched and iterated on agentic architectures, tool-calling LLM workflows (LangGraph) and deployed production MCP servers for standardized, authenticated, rate-limited tool access.

• Built internal AI copilots for customer-support triage, lead generation, and market research, integrating Slack, Calendly and other services to reduce manual workload by 90%

• Developed a SELF-DISCOVER method for long-form text style transfer, integrating contrastive style guides with LLMs, achieving a 44% improvement in target style accuracy and enabling human-editable guidelines for publication-grade content.

• Created a benchmark dataset of 600,000 semantically-paired documents across 5 domains (Tech, Entertainment, Finance, Food, Games), leveraging IoU scoring over Google keyword searches and TF-IDF for precise style differentiation.

• Designed and tested zero-shot, few-shot, and SELF-DISCOVER prompting strategies, incorporating BERT embeddings and Longformer models, boosting contextual precision by 30% in AI-driven writing tools.

• Trained classifiers on formality, politeness, humor, and simplicity, reducing stylistic deviation by 25% compared to baselines, as validated on a 100-pair benchmark subset.

• Led software development projects from 0 to 1, driving products from concept to production-ready state and guiding 40+ students through the development process.

• Collaborated with Dr. Neena Thota to streamline project mentorship, resolving complex technical challenges in real-time and implementing a rigorous grading rubric that improved student feedback loops by 20%.

• Engineered efficient ETL pipelines: Utilized Apache Kafka and Airbyte to integrate 14 data sources (PostgreSQL, REST APIs, and AWS S3 buckets). This reduced processing latency by 15%, enabling real-time data availability that powered machine learning models like Random Forests and LightGBM for actionable business insights.

• Crafted Python-based data parsing tools to transform unstructured datasets, cutting error rates by 28%

• Developed a recommendation system using collaborative filtering: Implemented the system with Apache Mahout, analyzing user behavior and purchase history to deliver tailored product suggestions.

• Implemented customer segmentation with K-means clustering: Applied scikit-learn to categorize users based on purchasing behavior and preferences, enabling precise targeting for marketing campaigns.