Engineering measurable, high-throughput AI for Science & scalable compute.
Data scientist and AI researcher building scientific AI systems with strong compute, modeling, and product execution. My work spans high-performance computing, representation learning, time-series forecasting, and quantum circuit knitting.
Featured Research & Scientific AI Systems
Click any paper card or prospectus button to open the comprehensive mathematical formulations, empirical benchmarks, and ablation studies.
Conference Presentations & Symposium Talks
Enhanced Intrusion Detection in IIoT Networks: A Lightweight Approach with Autoencoder-Based Feature Learning
Delivered oral presentation on deterministic cost-sensitive autoencoders deployed physically on NVIDIA Jetson Nano hardware for real-time Industrial IoT intrusion detection (99.94% Macro F1, 0.185 ms latency).
From Data to Action: Leveraging Machine Learning/Artificial Intelligence to Guide Proactive Pedestrian and Cyclist Safety Initiatives
Presented state-level machine learning frameworks and predictive spatial safety models for the Ohio Department of Transportation (ODOT) to evaluate high-risk road corridors, vulnerable road user crash severities, and countermeasure allocations.
Installation, Testing and Data Analysis for a Weather Station
Presented sensor deployment architectures, data telemetry collection pipelines, and statistical environmental analysis for meteorological monitoring systems supporting community resilience and extreme weather response.
Verified Peer Reviewer for Top-Tier Elsevier & IEEE Venues
Active referee service evaluating algorithmic rigor, theoretical soundness, and empirical validity across high-impact international artificial intelligence journals and conferences.
The builder behind the screen
AI for Science, self-supervised representation learning, large language models with vector retrieval validation, high-performance computing, and reproducible machine learning systems.
A model is only as good as its execution context. Designing algorithms that natively conform to memory hierarchies, SLURM batch constraints, and GPU interconnects is the difference between theoretical math and applied discovery.
Skills architecture
> torch.cuda.is_available() && sbatch --partition=ascend-gpu --nodes=4 --gpus=16
[OK] Slurm cluster allocation verified. 16x NVIDIA A100 Tensor Core GPUs linked via NVLink (600 GB/s).
Experience timeline
Formal Degrees & Institutional Credentials
M.S. in Computer Science & Engineering
Artificial Intelligence Track The University of Toledo · Toledo, Ohio, USA August 2023 – August 2025Core Focus: High-Performance Distributed Computing, AI for Science, Quantum Circuit Knitting, Parallel Model Partitioning.
B.Tech. in Electronics & Communication
Engineering Systems National Institute of Technology (NIT) · Srinagar, India July 2009 – July 2013Core Focus: Digital Signal Processing, Information & Communication Theory, Stochastic Modeling, Embedded Architectures.
35+ Professional Certifications & Coursework
In accordance with portfolio guidelines, the complete verified inventory of Stanford, Duke, UC Davis, and LinkedIn Learning certifications is hosted on its dedicated credentials page with live certificate thumbnail previews and on-demand modal loading.
Mufakir Qamar Ansari
Google Scholar verified profile connected for scheduled citation footprint and preprint metric ingestion. Focus domains include AI for Science, HPC acceleration, and quantum circuit knitting.
Open Source & Research Tooling
Full-stack implementations spanning distributed bioinformatics pipelines, quantum bisection schedulers, probabilistic day-ahead energy forecasting, and self-supervised medical image classifiers.
Let’s build serious AI systems.
Open to relocation within the United States. Actively seeking AI research appointments, high-throughput machine learning engineering roles, scientific computing collaborations, and high-leverage product engineering.