본 서비스에서 제공하는 사용자의 개인정보 데이터를 포함하여 각 정보주체의 동의 없이 데이터를 무단으로 수집하는 행위를 금지 및 거부합니다. 공개된 데이터도 크롤링 등 기술적 장치를 이용해 허가 없이 수집하는 경우 개인정보 보호법에 따라 형사처벌 될 수 있음을 알려드립니다.
© 2025 Rocketpunch, 주식회사 더블에이스, 홍원표, 대한민국 서울특별시 성동구 성수일로10길 12, 12층 1호, 04793, support@rocketpunch.com, +82 10-2710-7121
사업자등록번호 206-87-09615
더보기
타리크오메르
Ph.D. with 7 years of R&D experience specializing in privacy-preserving deep learning for federated edge computing, LLMs, and diffusion models, proficient in TensorFlow, PyTorch, ONNX & Python/C++.
경력
Assisting graduate students in course Special Topics in Computer Science (CS492).
더보기
– Developed Python scripts with Azure SDK and FastAPI to authenticate with Azure REST APIs and automate data ingestion from the Atlas platform, implementing real-time streaming pipelines for PRPD datasets using Apache Kafka. – Trained and optimized deep learning models using PyTorch and Hugging Face Transformers for fault diagnostic PRPD image classification, implementing quantization and TensorRT optimization for production deployment. – Implemented comprehensive MLOps pipeline using MLflow,Weights and Azure ML, with automated model versioning and deployment through GitHub Actions CI/CD workflows and containerized inference services using Docker and Kubernetes.
더보기
– Led SoC/RTL verification (SystemVerilog/UVM) for complex SoC designs integrating custom AI accelerators and secure data paths. – Developed optimized FPGA/SoC designs using Verilog, HLS (C/C++), and Vivado HLS for embedded AI applications, focusing on PPA (Power, Performance, Area) trade-offs. – Designed radiation-tolerant FPGA-based DSP systems for satellite communication, including signal processing accelerators for on-board AI-based image compression and packetization. – Developed imaging satellite payloads and high-speed multi-layer PCBs, implemented FPGA-based DSP/RF processing, and conducted Signal/Power Integrity analysis for high-frequency reliability. Awarded with: Distinguished Performance Award for PakTes-1A satellite launch.
더보기
학력
– Thesis: Robust Domain-Invariant Inertial Localization in Real-Time Federated Edge Computing – CGPA: 3.73/4.3 (94%) – Developed AIoT-enabled inertial localization, deep learning-based sensor fusion, and privacy-preserving navigation. – Developed adversarially robust and differentially private efficient deep-learning models for multimodal localization. – Designed NanoMST: an ultra-lightweight multiscale transformer for TinyML-based accurate inertial motion tracking. – Designed AccessBridge: an adaptive speech-to-text system with personalized ML models and cognitive accessibility features achieving 94.2% accuracy for assistive communication. – Benchmarked quantized neural networks for smartphone-based positioning in realtime federated edge computing. – Developed a real-time particle filter SLAM accelerator for mobile robot navigation. – Implemented end-to-end VSLAM with optimized feature-matching accelerators for real-time edge inference and mobile deployment.
더보기
