I’ve engineered and improved scalable real-time data pipelines and distributed systems for detecting and quarantining malicious emails, supporting hundreds of millions of messages daily with low-latency processing. Additionally, I lead the collaboration with our Data Science and ML team to build and improve our ML model lifecycle architecture.

• Architected and implemented Forta’s first cloud-native, microservices-based MLOps platform using Python, FastAPI, MLflow, Docker and Kubernetes, taking us from 0 to 20+ ML models, reducing model deployment time from weeks to minutes and enhancing the scalability of ML models in production
• Optimized Java and Python applications to reduce memory footprint and processing overhead by 60%, significantly boosting system reliability; leveraged tools like JProfiler, Pympler, Memray, and DataDog APM for in-depth memory profiling
• Enhanced multithreaded and asynchronous applications in an event-driven distributed architecture using Java’s ExecutorService, Python’s asyncio, and Amazon SQS; improved throughput and latency under high-concurrency workloads by optimizing thread pools, minimizing lock contention, and introducing batch processing
• Architected and developed robust, low-latency systems — such as a real-time campaign attack detection service in Python and FastAPI, leveraging Elasticsearch, Redis, and Amazon SQS — to enhance threat detection and response, while partnering with ML engineers to integrate ML models
• Enhanced and maintained a suite of metadata generation workflows using Apache Airflow and Amazon EMR with Spark, improving reliability and performance of downstream data pipelines
• Facilitated technical design discussions with product, data science, and customer success teams to solve technical issues, cutting incidents by 50% and boosting customer satisfaction
• Leveraged AWS Elasticache Redis as a transient data store, to reduce Kinesis data transfer costs by over 90% and eliminating data size constraints

Developed and maintained an end-to-end data streaming pipeline from the ground up, utilizing various tools to source, preprocess, map, and deliver large data sets, enabling actionable insights and effective compensation strategies

• Developed and maintained an end-to-end data streaming pipeline leveraging tools like Apache Hadoop, Spark, and Airflow, ensuring seamless data flow and real-time analytics capabilities
• Designed data warehouse architecture for new source systems, optimizing data storage and retrieval processes
• Created and implemented ETL tools, focusing on performance tuning and data transformation, while effectively managing high-priority tasks and adapting to evolving requirements

Delivered object-oriented software development for AT&T’s premier Software Defined Networking platform, focusing on the implementation and maintenance of network virtualization and automation features

• Developed and maintained software for AT&T’s Software Defined Networking platform, emphasizing network virtualization and automation capabilities
• Contributed to feature design, rigorous testing, and deployment processes to enhance platform functionality
• Integrated proof-of-concept solutions, ensuring rapid delivery through continuous integration and continuous deployment (CI/CD) pipelines