Building Intelligent & Scalable Data Systems
🔧 Currently: Data Scientist at Adept Tech Solutions — building production data modules and the rule-based generation engine for ASCEND, a synthetic data platform, alongside research on generative models for fraud detection
🎓 Education: B.E. Electrical Engineering from NUST
🔬 Research: Published at IEEE WCNC 2024 on Deep RL for resource allocation in MEC-enabled networks
🌱 Learning: Advanced Agentic AI systems and production ML deployment
⚡ Background: Came into ML from analog IC design — drawn to ML on hard physical problems
- ASCEND Synthetic Data Platform - Production data modules and the rule-based generation engine, plus research on generative models (GAN, autoregressive, cVAE) for fraud detection under severe class imbalance
- Job Intelligence Agent - Multi-agent system tracking job opportunities on a daily cron: REST/RSS/HTML collectors, dedup, sentence-transformer embeddings, two-layer ranking (cosine recall + LLM re-rank)
- MEC-NOMA Resource Allocation - Deep Deterministic Policy Gradient (DDPG) for computation offloading in IoT networks
- CNN-LSTM for Video-Based Regression - Video-based regression on the UBFC dataset with custom preprocessing and architectures
- ZeroPhish Gate - AI-powered phishing detection using hybrid BERT + LLaMA pipeline
- SAR ADC Design - 100 MHz 7-bit SAR ADC in TSMC 28nm PDK with bootstrapped sampling and dynamic latches
- RFID Transponder Chip - Ultra-low power passive RFID design for animal tagging in CMOS 65nm
- 4-bit Microprocessor - Custom FPGA design with VGA line drawing in Verilog
Optimizing Resource Allocation in MEC-Enabled CR-NOMA-Assisted IoT Networks: A DRL-Driven Strategy
IEEE Wireless Communications & Networking Conference (WCNC) 2024
Muhammad Taha Qaiser, Muhammad Sarmad Sohail, Minahil Shafqat, Syed Asad Ullah, Haejoon Jung, Syed Ali Hassan
Portfolio • LinkedIn • Google Scholar
"Building intelligent systems that scale"