Md Mashfiq Rizvee

Data engineer and Computer Science Ph.D. researcher with 4+ years of combined research and industry experience in data processing, database development, and machine learning workflows. Skilled in Python, SQL, PySpark, Spark SQL, and AWS, including EMR, S3, and SageMaker. Experienced in developing scalable ETL pipelines, transforming structured and unstructured data, assessing data quality, and preparing datasets for analytics and model evaluation.

Data Engineering Intern - Jumio Corporation - USA

(2026-04)

• Developed a production ready state-of-the-art DeepFake detection framework leveraging Fast Fourier Transform (FFT)-based frequency-domain artifacts to identify subtle inconsistencies introduced during synthetic face generation and manipulation. Deployed model in production achieved a 5% increase in accuracy in detecting deepfake transactions.
• Built and maintained ETL/ELT pipelines using Fivetran for source ingestion, Kafka for event streaming, and Databricks for data processing, delivering validated datasets to Snowflake. Automated workflows using Airflow.
• Extracted and analyzed authentication telemetry signals, including SDK versions, device characteristics, motion analysis, liveness scores, and photinus features, for model benchmarking and experimentation.
• Developed large-scale data querying pipelines using PySpark, Spark SQL, and AWS EMR to retrieve and analyze synthetic biometric authentication events for liveness fraud analysis.
• Developed and deployed data services on Azure using Docker, Kubernetes, and CI/CD workflows. Built knowledge graphs and GraphRAG pipelines to support contextual information retrieval.
• Collaborated on workflows supporting reward-informed model evaluation, verification-guided experimentation, and production training/evaluation pipelines, with applications in human authentication, anti-spoofing, multilingual/generalized robustness, and AI safety-oriented model testing.

Graduate Research Assistant - University of Kansas - USA

(2022-08)

• Developed efficient biometric database system designed to track humans using facial biometrics. Ensured enhanced security measures as well as established a comprehensive security framework for the database system, incorporating encryption, authentication, and access control mechanisms.
• Developed a secure database system designed to track a high volume of semiconductors within an electronic supply chain using PUFs. Evaluated results using SPC and DOE methods for monitoring variations over time. Collected PUF data from Ring-Oscillator-based RO-PUF with a MicroBlaze microprocessor deployed in the Basys-3 FPGA.
• Evaluated the security of biometric database systems by analyzing vulnerabilities against enrollment attacks, query attacks, denial-of-service (DoS) attacks, presentation attacks, and pre-image attacks. Documented threat models and mitigation strategies to strengthen authentication reliability across biometric and hardware-security deployments.
• Developed efficient CNN architectures and image classifiers that reduce the carbon footprint and energy consumption of AI systems. This work resulted in a patent.
• Collaborated with the Civil Engineering Department and developed a real-time road paving machine vision tool using OpenCV and FastSAM, enabling large CAT paver machines to detect paving boundaries and monitor asphalt distribution for construction safety. A pilot project was launched with the help of the City of Lawrence, Kansas.

Graduate Teaching Assistant - University of Kansas - USA

(2025-01)

• Instructed students through hands-on PCB soldering tasks in lab sessions, covering basic hardware assembly and circuit troubleshooting for digital circuit design principles.
• Guided students in using Xilinx Vivado for software-based labs involving logic building, simulation, synthesis, and implementation of digital circuits using VHDL and Verilog.

Doctor of Philosophy - Computer Science - University of Kansas (2025-05 - 2027-04)

Master of Science - Computer Science - University of Kansas (2022-08 - 2025-05)