Laeeq Aslam

Laeeq Aslam

GitHub LinkedIn Scholar PyPI

Aerial Wind Sensing | Wind-Field Reconstruction | Physics-Informed Learning

Postdoctoral Research Fellow · GTIIT
Contact Resume

Physics-Informed Aerial Wind Sensing and Wind-Field Reconstruction Postdoctoral research at GTIIT on aerial environmental sensing and flow-consistent learning

My current project at GTIIT develops low-cost multirotor drones as mobile wind-sensing platforms. The work estimates wind speed and direction from drone telemetry, attitude response, and reference calibration, then reconstructs local wind fields using Physics-Informed Neural Networks (PINNs) and flow-consistency constraints. In parallel, I work on physics-guided forecasting for wind-energy and grid-reliability systems, and climate-risk intelligence for urban thermal extremes.

Research Agenda

The central line is aerial wind sensing and wind-field reconstruction. Two parallel lines extend the same physics-guided learning philosophy to renewable-energy forecasting and urban climate-risk prediction.

SDG 7 Affordable and Clean Energy SDG 7 Affordable and Clean Energy SDG 11 Sustainable Cities and Communities SDG 11 Sustainable Cities and Communities SDG 13 Climate Action SDG 13 Climate Action

Current research trends in physics-informed AI for wind forecasting and urban climate resilience · AI for energy security, climate resilience, and infrastructure intelligence

Publications & Research

Recent work in physics-informed forecasting, aerial wind sensing, urban thermal risk, and embedded deployment.

Recent Publications

All Publications

Research

My research turns sparse, noisy field measurements into physically credible estimates of wind, energy-system behavior, and climate risk. The current GTIIT project is the main line; the other two areas run in parallel and support the same sensing-to-decision agenda.

Current GTIIT project

Low-cost drone-based wind estimation and wind-field reconstruction. I use multirotor telemetry, vehicle attitude, wind-relative direction, and reference calibration to infer ambient wind, then recover local flow structure with PINNs and flow-consistency constraints.

Parallel line in energy forecasting

Physics-guided forecasting for wind-energy and grid-reliability systems. I design short-horizon models that combine temporal learning, physical transport structure, and regime-aware correction under ramps, extremes, and site-dependent variability.

Parallel line in climate risk

Climate-risk intelligence for urban heat and infrastructure resilience. I develop physics-informed spatio-temporal models that preserve spatial context, physical plausibility, and useful lead time for intervention-oriented forecasting.

Projects

Selected projects organized around the main postdoctoral direction and two parallel research lines.

Current project at GTIIT

Low-cost drone wind estimation

Develops calibration and learning methods that estimate wind speed and direction from low-cost multirotor telemetry, attitude response, and reference measurements during real outdoor flights.

Current project at GTIIT

PINN-based wind-field reconstruction

Reconstructs local wind fields from sparse aerial observations using Physics-Informed Neural Networks, flow regularization, and consistency constraints that encode physically plausible wind evolution.

Parallel research line

Physics-guided wind and grid forecasting

Couples adaptive feature selection, physical wind-field structure, and sequence learning to improve short-horizon prediction for renewable-energy integration and grid reliability.

Parallel research line

Urban heat and climate-risk intelligence

Uses physics-informed ensemble learning and residual spatio-temporal correction to forecast city-center thermal extremes and support infrastructure-resilience decisions.

News

Recent academic updates.

  1. 2026 Joined Guangdong Technion–Israel Institute of Technology as a Postdoctoral Research Fellow.
  2. 2025 Published physics-informed ensemble learning work for thermal extremes in Urban Climate.
  3. 2025 Published physics-informed spatio-temporal wind forecasting work in Computers & Electrical Engineering.
  4. 2025 Published edge-oriented wind prediction work for memory-constrained devices.

Experience

Track record from embedded systems and high-throughput vision pipelines to physics-informed forecasting for energy security, climate resilience, and constrained hardware deployment.

Postdoctoral Research Fellow

Guangdong Technion–Israel Institute of Technology (GTIIT) · Mar 2026 – Present

Current

I work on low-cost drone-based wind estimation and wind-field reconstruction. This research uses multirotor telemetry, reference wind measurements, and physics-informed learning to estimate ambient wind and recover local flow structure from sparse aerial observations.

The current focus is robust estimation under real flight conditions, Physics-Informed Neural Network reconstruction, cross-platform generalization, and deployable inference pipelines for field experiments and atmospheric sensing.

PhD Researcher – Physics-Informed AI for Energy and Climate Systems

Central South University · Sept 2020 – Dec 2025

PhD

I design physics-informed sequence models for wind, load, and near-surface temperature forecasting, targeting grid reliability and urban thermal-stress early warning. This includes regime-adaptive focal losses, adaptive physics penalties, attention-equipped hybrid recurrent / convolutional temporal modules, and hardware-aware optimization for Jetson-class devices.

I led research end-to-end from problem formulation to publication, producing 6 first-author papers. The work aligns with a provincial Key R&D program on wind energy and grid optimization (Project 2020WK2007).

AI/ML Consultant – Edge Computing & Computer Vision

Bond and Built Pvt Ltd · Jul 2024 – Mar 2025

Industry

Built and scaled a footwear analytics pipeline handling 60+ concurrent video streams and ~50,000 daily inferences. Delivered production-grade market intelligence under real deployment and bandwidth constraints.

Machine Learning Engineer – Vision & Deployment

DLISION · May 2021 – Sept 2022

Industry

Led semantic segmentation and real-time detection pipelines. Tuned UNet variants with attention and focal loss to improve segmentation accuracy by ~3% over in-house baselines and optimized inference for deployment.

Lecturer & Research Assistant – Embedded Systems / AI

iUSE School of Engineering & IIUI · 2013 – 2020

Taught embedded systems, programming, and discrete-time signal processing. Built an Instrumentation & Measurement Lab for sensor-driven data acquisition in MATLAB. Supervised applied student projects in energy management, safety telemetry, and secure voting.

Investigated GAN-based medical diagnostics for breast cancer, with emphasis on data augmentation and robustness.

Education & Skills

Education

  1. PhD in Control Science & Machine Learning

    Central South University, China

    Sept 2020 – Dec 2025

  2. MS in Electronic Engineering Gold Medalist

    International Islamic University, Pakistan

    Sept 2015 – Aug 2017

  3. BS in Electronic Engineering

    International Islamic University, Pakistan

    Sept 2006 – Aug 2010

Technical Skills

Machine Learning & AI
Deep Learning · Computer Vision · Time-Series Forecasting
Programming & Tools
Python · MATLAB · C++ · TensorFlow · PyTorch · OpenCV · Edge Impulse
Deployment & Systems
Docker · AWS (EC2) · Git · Triton Inference Server
Hardware & IoT
Raspberry Pi · NVIDIA Jetson · Arduino Nano BLE Sense
Data & Optimization
Pandas · NumPy · Matplotlib · Seaborn · Feature Engineering

Awards & Memberships

Chinese Government CSC Scholarship (PhD)
Hunan Provincial Key R&D Program (Project 2020WK2007) — wind energy & grid optimization
Gold Medalist — MS in Electronic Engineering
Registered Engineer, Pakistan Engineering Council (ELECTRO/22837)
Member, IEEE

Research Tooling

Lightweight, deployment-aware tooling. Built for reproducibility, ablation, and on-device inference.

ed-rvfl-sc

PyPI CPU-only Edge-ready

MIT License · Python ≥3.9

Repo

Ensemble Deep Random Vector Functional Link with skip connections (edRVFL-SC). Delivers deep-ML–level accuracy without GPU training, using closed-form layer solves and feature reuse. Targets ultra-fast training and inference on CPU and embedded boards.

Install
pip install ed-rvfl-sc
Non-iterative training,
100× faster setup,
Jetson-friendly.

TimeMesh

PyPI Forecasting pipeline

PyTorch · Temporal CNNs / LSTM / Transformer

Repo

Preprocessing and dataset tooling for time-series forecasting experiments. Sliding windows, normalization flows, train/val/test slicing, and ready-to-train tensors for LSTM, TCN, Transformer, and iTransformer-style models.

Install
pip install timemesh
Reproducible splits,
clean dataloaders,
fast ablations.

Keras Swin-UNet

PyPI Segmentation

TensorFlow · Vision Transformer · Focal Loss

Repo

Transformer-based Swin-UNet segmentation stack for earth observation, medical imaging, and industrial perception. Includes attention backbones and robust Focal Loss settings for rare-structure segmentation.

Install
pip install keras-swin-unet
Strong baselines,
clean API,
industry-friendly.

Contact

I welcome research correspondence on low-cost drone-based wind estimation, PINN-based wind-field reconstruction, physics-guided energy forecasting, and climate-risk intelligence.

Email

laeeq.aslam.100@gmail.com

Guangdong Technion–Israel Institute of Technology (GTIIT), Shantou, Guangdong, China.

Profiles

Google Scholar GitHub LinkedIn PyPI