Machine Learning Engineer with 3+ years of experience building scalable AI systems, specializing in Generative AI, Large Language Models (LLMs), and Retrieval-Augmented Generation (RAG). Proven track record of developing…
Machine Learning Engineer with 3+ years of experience building scalable AI systems, specializing in Generative AI, Large Language Models (LLMs), and Retrieval-Augmented Generation (RAG). Proven track record of developing real-time ML inference platforms, vector search systems, and distributed ML pipelines using Python, PyTorch, Spark, and Kubernetes. Experienced in MLOps, cloud-native deployments, and event-driven architectures on AWS for high-performance AI applications. Passionate about designing production-grade AI platforms powering enterprise assistants, personalization engines, and intelligent decision systems.
Engineered a production-scale Retrieval-Augmented Generation (RAG) architecture powering enterprise AI assistants, reducing hallucination rates by 35% and improving response accuracy across knowledge-base queries using LLM integrations from OpenAI and Anthropic.
Built real-time LLM inference services deployed on GPU Kubernetes clusters, achieving sub-500ms response latency and enabling high-throughput AI copilots supporting thousands of concurrent enterprise users.
Designed an AI personalization and recommendation engine leveraging user embeddings and ranking models, increasing contextual response relevance by 40% and enabling targeted content and ad-based monetization workflows.
Developed distributed ML pipelines using Python, PyTorch, Ray, and Apache Spark to process large-scale multimodal datasets (text, image, and video) for model training and evaluation.
Built scalable vector retrieval systems using FAISS, Pinecone, and embedding models, enabling semantic search across millions of documents for enterprise AI applications.
Implemented LLM orchestration microservices with FastAPI, Docker, and Kubernetes, integrating APIs from Meta Llama models and third-party model providers to support multi-model routing and prompt optimization.
Designed event-driven AI data pipelines using Apache Kafka, Airflow, and cloud storage on Amazon Web Services, enabling automated dataset ingestion, labeling workflows, and model evaluation at scale.
Engineered a production-scale Retrieval-Augmented Generation (RAG) architecture powering enterprise AI assistants, reducing hallucination rates by 35% and improving response accuracy across knowledge-base queries using LLM integrations from OpenAI and Anthropic., Built real-time LLM inference services deployed on GPU Kubernetes clusters, achieving sub-500ms response latency and enabling high-throughput AI copilots supporting thousands of concurrent enterprise users., Designed an AI personalization and recommendation engine leveraging user embeddings and ranking models, increasing contextual response relevance by 40% and enabling targeted content and ad-based monetization workflows., Developed distributed ML pipelines using Python, PyTorch, Ray, and Apache Spark to process large-scale multimodal datasets (text, image, and video) for model training and evaluation., Built scalable vector retrieval systems using FAISS, Pinecone, and embedding models, enabling semantic search across millions of documents for enterprise AI applications., Implemented LLM orchestration microservices with FastAPI, Docker, and Kubernetes, integrating APIs from Meta Llama models and third-party model providers to support multi-model routing and prompt optimization., Designed event-driven AI data pipelines using Apache Kafka, Airflow, and cloud storage on Amazon Web Services, enabling automated dataset ingestion, labeling workflows, and model evaluation at scale.
Designed and deployed machine learning credit risk models using Python, XGBoost, and Scikit-learn to automate loan underwriting decisions, improving credit risk prediction accuracy by ~18% and reducing manual review workload across the digital lending pipeline.
Architected a real-time AI inference system using Kafka streaming, FastAPI microservices, and Kubernetes, enabling sub-second loan eligibility scoring and accelerating loan approval processing time by 35%.
Built scalable feature engineering pipelines using Apache Spark, PySpark, and SQL to process high-volume financial and behavioral data, enabling consistent feature generation for credit risk, fraud detection, and personalization models.
Implemented MLOps pipelines with MLflow, Kubeflow, and Docker to automate model training, versioning, and deployment, improving reproducibility and reducing model release cycles across cloud environments.
Developed document AI pipelines for automated loan document verification using OCR, NLP entity extraction, and Python-based preprocessing, converting unstructured financial documents into structured data for downstream ML decision engines.
Deployed containerized ML models on cloud-native infrastructure (Kubernetes, REST APIs, and scalable microservices) while integrating real-time monitoring using Prometheus and Grafana to track model performance, latency, and drift in production systems.
Designed and deployed machine learning credit risk models using Python, XGBoost, and Scikit-learn to automate loan underwriting decisions, improving credit risk prediction accuracy by ~18% and reducing manual review workload across the digital lending pipeline., Architected a real-time AI inference system using Kafka streaming, FastAPI microservices, and Kubernetes, enabling sub-second loan eligibility scoring and accelerating loan approval processing time by 35%., Built scalable feature engineering pipelines using Apache Spark, PySpark, and SQL to process high-volume financial and behavioral data, enabling consistent feature generation for credit risk, fraud detection, and personalization models., Implemented MLOps pipelines with MLflow, Kubeflow, and Docker to automate model training, versioning, and deployment, improving reproducibility and reducing model release cycles across cloud environments., Developed document AI pipelines for automated loan document verification using OCR, NLP entity extraction, and Python-based preprocessing, converting unstructured financial documents into structured data for downstream ML decision engines., Deployed containerized ML models on cloud-native infrastructure (Kubernetes, REST APIs, and scalable microservices) while integrating real-time monitoring using Prometheus and Grafana to track model performance, latency, and drift in production systems.
The AI-first Recruitment Operating System that unifies and automates the entire hiring lifecycle for modern enterprises worldwide.
Dr Job FZ LLC. 2026 © All Rights Reserved
