AI/Backend Software Engineer at Ampliv(a)

Modernized and scaled payment and ML platforms. Decomposed a Flask payments monolith into eight FastAPI microservices behind a Django JWT authentication gateway, reducing p95 latency by 40% and doubling peak throughput. Engineered a Kafka-driven fraud detection pipeline using Python and TensorFlow, reducing false positives by 30% while adding GDPR-compliant audit logging and traceability. Processed and optimized transaction datasets with Pandas and SQL for ML training, improving model precision by 15%. Designed an LLM-powered semantic search solution over transaction logs using LangChain and OpenAI embeddings to accelerate fraud investigation workflows. Developed and deployed real-time NLP inference APIs with FastAPI and Docker on AWS ECS, supporting 2,000+ requests per second with sub-100ms response latency.

AI/Backend Software Engineer at Ampliva

Modernized a Flask payments monolith by decomposing it into eight FastAPI microservices behind a Django JWT authentication gateway, reducing p95 latency by 40% and increasing peak throughput by 2x. Engineered a Kafka-driven fraud detection pipeline using Python and TensorFlow, decreasing false positives by 30% and implementing GDPR-compliant audit logging and traceability. Processed and optimized transaction datasets with Pandas and SQL for model training, improving model precision by 15%. Designed and prototyped an LLM-powered semantic search solution for transaction logs using LangChain and OpenAI embeddings to accelerate fraud investigation and analysis. Developed and deployed real-time NLP inference APIs with FastAPI and Docker on AWS ECS, supporting over 2,000 requests per second with sub-100ms latency.

Senior Software Engineer - AI/ML at NALA

Consolidated multiple Magento REST APIs through an Apollo GraphQL gateway built on Django ASGI, reducing response payload sizes by 30% and improving feature development speed by 25%. Developed automated inventory analytics workflows using MongoDB, Pandas, and Power BI, cutting weekly reporting time from 5 hours to 2 and adding nightly data validation. Integrated an NLP proof of concept using spaCy and SentenceTransformers for customer review analysis and product insight extraction, increasing recommendation accuracy by 12%. Built and tested LangChain-based AI agents to create an internal chatbot for inventory and pricing inquiries, reducing manual data retrieval efforts by 40%.

Software Developer at Truvio

Developed an asynchronous FastAPI-based purchase order processing service with Celery workers, reducing reconciliation timelines from three days to same-day accuracy. Built a lightweight Express middleware layer for ERP integration and implemented spaCy-based NLP automation to extract data from supplier emails and generate purchase orders, reducing manual data entry by 85%. Implemented real-time inventory monitoring using WebSocket alerts and Slack slash commands, keeping stock-out rates below 2% across 1.8K SKUs. Designed and implemented a multi-stage CI/CD pipeline with Docker, Trivy, and semantic-release to automate testing, security scanning, and weekly deployment of ML prototypes and API services.