Danilo Brontolone

– Senior Backend/AI Engineer on a 3-person core team building an AI-powered healthcare intake platform. Led design and implementation of multi-agent orchestration, context-aware session memory, and RAG pipelines, delivering scalable and reliable AI workflows with voice and chat integration. – Architected and deployed multi-agent AI orchestration systems, automating complex healthcare workflows and improving end-to-end task efficiency by 15-20%. – Designed and optimized Retrieval-Augmented Generation (RAG) pipelines, increasing document retrieval precision by 10-15% and saving 5-10 hours of manual lookup per week. – Developed and productionized an AI voice agent handling 50-100 calls/day, achieving ~85% intent recognition accuracy and reducing human agent workload by 15-20%. – Implemented contextual memory and multi-turn reasoning, boosting conversation retention and task completion rates from ~60% to ~70%. – Deployed and maintained scalable, fault-tolerant AI services using Docker, Kubernetes, and AWS, consistently achieving >99% service uptime.

– AI Engineer in a fast-paced agency setting building production RFP proposal automation and AI document structuring systems. Owned end-to-end design of multi-agent LLM workflows and document intelligence pipelines across multiple client engagements. – Developed and scaled conversational AI platforms serving 2,000–5,000 monthly active users, improving average response accuracy by ~15% through iterative model tuning and evaluation. – Built AI-driven analytics and reporting pipelines, reducing report generation cycles from ~3 days to 6-8 hours, enabling faster business decision-making. – Designed and executed prompt engineering strategies and LLM evaluation workflows, increasing response relevance and consistency by 10-15% across production use cases. – Implemented text embedding and semantic search pipelines, improving document retrieval precision from ~65% to ~75% through optimized chunking and similarity scoring. – Integrated AI capabilities into production-grade REST and GraphQL APIs, reducing end-to-end API latency by 10-15% while maintaining reliability and scalability.

– Developed and optimized machine learning models, improving prediction accuracy across client pipelines by 10–15% through feature engineering and model tuning. – Built and deployed RESTful inference APIs serving ~1,000-2,000 requests per day, reducing average client response times by ~20%. – Streamlined and automated data ingestion and ETL pipelines, cutting end-to-end processing time by 15–20% and improving data reliability. – Delivered two end-to-end ML-powered applications, from data preparation and modeling to deployment, meeting all project timelines and requirements. – Implemented CI/CD pipelines for ML services, reducing deployment-related errors by ~30% and accelerating release cycles.

– Trained and evaluated neural network-based NLP models, improving classification accuracy by ~5–10% through iterative experimentation and hyperparameter tuning. – Processed, cleaned, and labeled 1M+ text records to support large-scale model training and evaluation. – Developed supervised and unsupervised learning pipelines, reducing feature extraction and preprocessing time by ~15%. – Contributed to the development of two research-grade ML prototypes, later validated through pilot deployments. – Performed data preprocessing, feature engineering, and model evaluation to systematically improve model performance and stability. – Collaborated with cross-functional teams (product, engineering, research) to gather requirements, align on objectives, and support solution delivery.