PhD Researcher & Engineer
Muhammad Umair, PhD
AI Researcher | AI Solution Engineer
I design and deploy AI systems that bridge academic research and real-world impact.
Connect With Me
PhD Researcher & Engineer
I design and deploy AI systems that bridge academic research and real-world impact.
Profile
Researcher and AI engineer focused on multimodal AI, cross-modal alignment, and robust generalization.
My work focuses on multimodal AI, cross-modal learning, and efficient model design for real-world deployment. I completed my PhD in Information and Communication Engineering at the University of Electronic Science and Technology of China (UESTC), under the supervision of Prof. Zhou Jun (Team leader IoT Smart Chip and System).
My PhD thesis, Multimodal Person Re-Identification for Low-Resource Regimes, explored identity matching across RGB, thermal infrared, infrared, and text modalities, with an emphasis on robust generalization, cross-modal alignment, and computational efficiency.
A central highlight of my doctoral work was the creation of a large-scale multimodal person re-identification dataset & multi-model covering RGB, thermal infrared, infrared, and text, complemented by methods for alignment and robust test-time adaptation.
In industry and public-sector collaborations, I have delivered 50+ applied AI projects for UNDP, I‑SAPS, LAS Pakistan, and companies including ESGnie, VikkAI, Lincware, BlueOcean, Staff, Cracked.AI, and Atomcamp.
I also lead professional training in Generative AI, Agentic AI, NLP, and MLOps for national and international audiences.
University of Electronic Science and Technology of China
National Textile University, Faisalabad, Pakistan
University of Agriculture, Faisalabad, Pakistan
Focus Areas
Alignment across vision-language modalities, including RGB, thermal, infrared, and text.
Identity matching, retrieval, and benchmarking for multimodal re-identification.
Domain generalization and test-time adaptation for reliable deployment.
Reproducible experiments, ablation studies, and evaluation protocol design.
Selected
CVPR 2026 Workshop SynData4CV , 2026 | Conference Paper
Information Processing & Management (Zone 1, Q1), 2026 | Journal Article
ICCC 2025 | Conference Paper
International Conference on Engineering Software for Modern Challenges, 2021 | Conference Paper
Neural Computing and Applications (Q1), 2023 | Journal Article
Future Networks and Distributed Systems, 2018 | Conference Paper
ICIC 2021 | Conference Paper
Engagements
End-to-end research and engineering support for multimodal systems, evaluation design, and AI deployment.
Expert guidance on implementing AI solutions in your organization, including technology selection, architecture design, and ROI analysis.
Development of custom AI models and systems tailored to your specific requirements and business objectives.
Comprehensive training programs ranging from introductory workshops to advanced technical sessions for your team.
Appointments
Selected roles across research, consulting, and applied AI engineering.
Current
Advising on AI product strategy and deployment for enterprise and startup partners.
Current
Designed and delivered programs in Generative AI, Agentic AI, NLP, and MLOps.
University of Lahore
Teaches AI & NLP.
June 2022 to March 2023
Delivered NLP models, production pipelines, and cloud deployments on GCP.
Selected partners and collaborations:
Selected Work
Applied research systems spanning multimodal perception, language understanding, and cross-modal retrieval.
Built a large-scale benchmark spanning RGB, thermal infrared, infrared, and text with alignment-focused evaluation.
Assessed Kolmogorov-Arnold Transformer based methods for object and person re-identification under practical constraints.
Led analysis of public sentiment around social and religious events for UNDP and LAS Pakistan.
Recognition
Selected awards, distinctions, and professional recognition.
2024 Graduate Academic Forum, SICE, UESTC
Awarded with a 2,000 元 prize, supported by the Huang Shunji Education Development Fund.
TCM Originals
Recognized for applied AI expertise and public engagement.
IBM (Coursera)
Professional credential focused on collaboration and leadership.
Contact
Interested in collaborating or learning more about my work? Reach out for consulting inquiries, research collaborations, or training opportunities.
muhammadumair894@gmail.com
Chengdu, China
Book a meeting via Google Calendar
