Syed Asnal Rizvi

Directed Transak’s AI-assisted Risk Engine, examining behavioral and transactional signals for real-time fraud detection. Architecting scalable, modular backend solutions supporting high-throughput financial transactions and AI workloads. Monitoring and optimizing backend performance, latency, and database efficiency for mission-critical applications. Developed adaptive scoring models that boosted fraud precision and approval rates by 15%, averting $400K+ in annual chargeback losses.

Engineered a robust and scalable backend for Rene from the ground up using Python, MySQL, and ElasticSearch. Executed a combination of K-means clustering, NLTK, and BERT with Python and TensorFlow on Enron’s email dataset to classify messages and categorized them into labels such as sales, hiring, follow-up, and more.

Architected Python-based payments back-end systems, driving a 30% ARR growth (~$360K) via fiat and crypto integrations. Enhanced system reliability and throughput for 100K+ users by streamlining microservice orchestration.

Developed GenAI-powered content generation tools using Python, reducing writing time by 40% and proofreading effort by 70%, achieving $300K+ in annual savings. Implemented NLP pipelines for automated data extraction, enhancing research speed and accuracy.

Designed an AI-powered conversational system (AERIN) to assist users with itinerary planning using Python, NLP & Elasticsearch. Accelerated response times from 6s to under 2s, driving a 25% increase in user conversions. Expanded system architecture to support 1,000+ concurrent sessions seamlessly.