AI Engineer with 4+ years of experience designing and deploying scalable machine learning and generative AI systems across fintech and enterprise environments. Skilled in LLM-powered applications, Retrieval-Augmented Gen…
AI Engineer with 4+ years of experience designing and deploying scalable machine learning and generative AI systems across fintech and enterprise environments. Skilled in LLM-powered applications, Retrieval-Augmented Generation (RAG), real-time fraud detection, anomaly detection, and predictive analytics using cloud-native MLOps pipelines. Experienced in building end-to-end AI workflows using PyTorch, TensorFlow, Spark, MLflow, Kubernetes, and Vertex AI, with strong expertise in semantic search, embedding models, distributed data processing, and production-grade ML infrastructure. Proven ability to operationalize AI systems at scale, improve decision-making through intelligent automation, and support high-volume real-time inference workloads.
Designed LLM-powered fraud risk classification system for digital banking transactions using Retrieval-Augmented Generation and prompt engineering, improving suspicious activity detection precision by 19% across real-time payment workflows.
Developed generative AI pipeline using Hugging Face Transformers to automate dispute resolution summaries, reducing customer support handling time by 23% across peer-to-peer payment service operations.
Implemented vector similarity search using embedding models and FAISS-based vector database to enhance semantic transaction categorization accuracy across user financial activity logs in large-scale consumer banking datasets.
Deployed containerized generative inference pipelines using Docker and Kubeflow, enabling scalable AI-based anomaly insights across account behavior monitoring systems within cloud-hosted fintech platforms.
Integrated experiment tracking and model governance using DVC to support controlled A/B testing for AI-driven spending pattern insights across mobile-first digital banking applications.
Built cloud-native deployment workflows in Google Vertex AI for real-time credit risk scoring models, processing over 3 million monthly user transactions across embedded financial service ecosystems.
Designed LLM-powered fraud risk classification system for digital banking transactions using Retrieval-Augmented Generation and prompt engineering, improving suspicious activity detection precision by 19% across real-time payment workflows., Developed generative AI pipeline using Hugging Face Transformers to automate dispute resolution summaries, reducing customer support handling time by 23% across peer-to-peer payment service operations., Implemented vector similarity search using embedding models and FAISS-based vector database to enhance semantic transaction categorization accuracy across user financial activity logs in large-scale consumer banking datasets., Deployed containerized generative inference pipelines using Docker and Kubeflow, enabling scalable AI-based anomaly insights across account behavior monitoring systems within cloud-hosted fintech platforms., Integrated experiment tracking and model governance using DVC to support controlled A/B testing for AI-driven spending pattern insights across mobile-first digital banking applications., Built cloud-native deployment workflows in Google Vertex AI for real-time credit risk scoring models, processing over 3 million monthly user transactions across embedded financial service ecosystems.
Built demand forecasting pipelines for a retail supply chain client using PySpark and LightGBM, increasing weekly inventory planning accuracy by 21% across geographically distributed warehouse fulfillment networks.
Architected Databricks-based ETL workflows to process over 25 million transactional records, enabling stable feature engineering for downstream predictive analytics across seasonal and promotional retail demand fluctuations.
Deployed anomaly detection models using Scikit-learn to monitor supplier delivery inconsistencies, decreasing order fulfillment cycle delays by 17% across logistics operations in tier-2 regional distribution hubs.
Implemented semantic document retrieval using embedding models with RAG architecture, improving internal support ticket resolution turnaround time by 26% across customer assistance platforms.
Operationalized MLflow-integrated monitoring pipelines on Kubernetes to track deployment performance, supporting consistent experimentation across multiple production-ready forecasting models within AWS SageMaker runtime environments.
Enabled real-time inference through FastAPI-based RESTful services, facilitating automated stock replenishment recommendations across online retail applications during peak transaction intervals without service degradation.
Designed centralized feature store architecture in Snowflake for dataset versioning, improving model training reproducibility and supporting governed segmentation workflows across omnichannel customer engagement initiatives.
Built demand forecasting pipelines for a retail supply chain client using PySpark and LightGBM, increasing weekly inventory planning accuracy by 21% across geographically distributed warehouse fulfillment networks., Architected Databricks-based ETL workflows to process over 25 million transactional records, enabling stable feature engineering for downstream predictive analytics across seasonal and promotional retail demand fluctuations., Deployed anomaly detection models using Scikit-learn to monitor supplier delivery inconsistencies, decreasing order fulfillment cycle delays by 17% across logistics operations in tier-2 regional distribution hubs., Implemented semantic document retrieval using embedding models with RAG architecture, improving internal support ticket resolution turnaround time by 26% across customer assistance platforms., Operationalized MLflow-integrated monitoring pipelines on Kubernetes to track deployment performance, supporting consistent experimentation across multiple production-ready forecasting models within AWS SageMaker runtime environments., Enabled real-time inference through FastAPI-based RESTful services, facilitating automated stock replenishment recommendations across online retail applications during peak transaction intervals without service degradation., Designed centralized feature store architecture in Snowflake for dataset versioning, improving model training reproducibility and supporting governed segmentation workflows across omnichannel customer engagement initiatives.
Engineered AI-powered smart switchboard integrating ESP8266 relays, Raspberry Pi edge intelligence, and Flutter mobile control, enabling remote automation and reducing manual device management across connected homes. Developed QR-based onboarding system that automatically registers new switchboards via Firebase and MQTT, simplifying installation for users and eliminating manual configuration during multi-room deployments efficiently overall. Implemented edge AI models on Raspberry Pi to analyze sensor data and predict energy usage, enabling automated device scheduling and reducing unnecessary power consumption across households. Designed resilient MQTT-based communication architecture supporting offline operation, buffering device events locally and synchronizing with Firebase when connectivity returns, ensuring uninterrupted automation and monitoring for users. Integrated motion, light, and temperature sensors with anomaly detection algorithms to trigger intelligent automation routines and security alerts, improving safety while optimizing household energy utilization daily.
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
