Boris S.

Work experience

• Senior AI/ML Engineer – Agentic AI & LLMsFeb 2025 - May 20261 y.

  Architected hybrid Retrieval-Augmented Generation (RAG) pipelines integrating Pinecone, FAISS, TinyBERT, and Hugging Face Transformers to power medical knowledge retrieval, improving diagnostic accuracy by 38% and reducing hallucinations by 60%.

  Responsibilities:

  • Designed multi-modal reasoning pipelines combining structured clinical data, medical imaging, and unstructured text using GraphRAG and Neo4j knowledge graphs.
  • Built reinforcement learning pipelines (RLHF, PPO, GRPO) enabling continuous optimization of treatment recommendations based on clinician and patient feedback.
  • Developed a high-throughput LLM inference platform using AWS EKS, SageMaker, TensorRT-LLM, and FlashAttention, reducing model latency by 55% while supporting thousands of concurrent queries.
  • Implemented end-to-end MLOps pipelines including CI/CD automation, model monitoring, and retraining workflows to ensure production reliability in regulated healthcare environments.
  • Led cross-functional initiatives to develop multi-agent AI orchestration systems, enabling collaboration between diagnostic models, predictive analytics modules, and recommendation engines.
• Senior AI/ML EngineerJan 2023 - Jan 20252 y.

  Designed enterprise-grade ML platforms on a company supporting large-scale predictive analytics and recommendation systems across multi-terabyte datasets.

  Responsibilities:

  • Developed deep learning models using PyTorch and TensorFlow across tabular, text, and image modalities, improving predictive accuracy by 25%.
  • Built scalable MLOps infrastructure with Kubeflow, MLflow, Docker, and Kubernetes, reducing model deployment cycles and improving production reliability.
  • Optimized distributed Spark feature engineering pipelines, improving ETL performance by 40%.
  • Partnered with product and data teams to integrate AI capabilities into enterprise applications used by 50,000+ users.
  • Mentored junior engineers and led architecture reviews to improve code quality and ML deployment standards.
• Senior Machine Learning EngineerMar 2021 - Nov 20222 y.

  Built NLP and computer vision models using transformer architectures (BERT, GPT) to automate enterprise document processing and visual intelligence workflows.

  Responsibilities:

  • Designed real-time inference APIs using FastAPI, Docker, and Kubernetes capable of processing millions of requests monthly.
  • Developed end-to-end ML pipelines with Airflow, Spark, and MLflow, reducing model-to-production time by 30%.
  • Improved NLP model performance 20-25% through ensemble methods, knowledge distillation, and transfer learning.
  • Conducted research on multi-agent generative AI systems using RAG architectures and Claude-based LLM agents.
• Machine Learning EngineerJul 2018 - Feb 20213 y.

  Developed ML solutions for fraud detection, churn prediction, and recommendation systems for fintech and e-commerce clients.

  Responsibilities:

  • Built scalable data pipelines using Python, Pandas, and Spark for large-scale transactional data processing.
  • Trained and optimized models using XGBoost, LightGBM, and scikit-learn to improve fraud detection precision and reduce false positives.
  • Deployed ML models as REST APIs using Flask and Docker integrated with enterprise analytics systems.
  • Built batch inference pipelines on AWS processing millions of predictions daily.
• AI Engineer / Data ScientistJun 2014 - Jun 20184 y.

  Developed predictive models including regression, classification, and clustering algorithms for business intelligence and forecasting.

  Responsibilities:

  • Built data processing pipelines using Python, SQL, and Pandas to extract insights from structured and unstructured datasets.
  • Implemented early deep learning solutions including variational autoencoders and neural networks for automation and predictive analytics.
  • Deployed ML services on AWS using Docker-based environments to enable scalable inference pipelines.