HKPFS Fellow PolyU PhD (2026) McGill Visiting Researcher

Making Infrastructure Condition Assessment Measurable, Auditable & AI-Driven

I build systems that transform subjective civil engineering inspections into auditable numbers. Combining deep learning, multi-sensor data fusion—Ground-Penetrating Radar, acoustics, UAV imagery, IoT telemetry—and LLM decision-support platforms into production-grade software engineers can run.

0
Peer-Reviewed Papers
0
Journal Articles
0
Q1 Top Journals
0
+ Citations
0
h-index

Researcher & System Architect for Civil Assets

Current Appointment

Postdoctoral Fellow

Department of Construction Management and Innovation

The Hong Kong Polytechnic University (PolyU)

Doctoral Distinction

PhD in Civil Engineering (2026)

Awarded under the prestigious Hong Kong PhD Fellowship Scheme (HKPFS).

Dissertation: Integrated Pavement Performance Assessment System Using Multi-Sensing Technologies.

Global Academic Mobility

Postgraduate Visiting Researcher

Department of Civil Engineering & Applied Mechanics, McGill University, Canada.

Winter road surface condition modeling, remote sensing analytics, and Multimodal LLM integration for climate resilience.

What I Work On

Bridging computer vision, non-destructive sensing, LLM synthetic experts, sustainable materials, and mathematical optimization for built environment infrastructure.

Automated Condition Assessment

Segmentation and detection models for cracks, rutting, and surface distress (SegFormer, FPN, YOLOv8, SAM 3). Every prediction is strictly validated against measured ground truth rather than eyeballed visual estimates.

PyTorch SegFormer SAM 3 YOLOv8 Diffusion AI500

Subsurface & Non-Destructive Sensing

High-frequency Ground-Penetrating Radar (GPR) signal processing, layer interfacing, pavement thickness evaluation, bridge-deck rebar recognition, cover depth measurement, and early-stage corrosion assessment.

GPR Signal Analytics B-Scan Interfacing Cover Depth Rebar Corrosion

Utility Management Using Sensing Technologies

AIoT data pipelines and machine learning algorithms operating on hydrophone and noise-logger telemetry to detect, classify, and pinpoint leaks in live urban water distribution networks under operational noise, as well as AI-driven sewer condition assessment and defect recognition.

Hydrophones AIoT Telemetry Sewer Defect Detection Water & Wastewater Networks

Decision Support & LLMs

Zero-shot Multimodal Large Language Models (MLLMs) acting as synthetic experts for visual distress evaluation, fuzzy AHP, deductive indices, and LLM-assisted maintenance prioritization platforms.

Multimodal LLMs Synthetic Experts Fuzzy AHP Prompt Engineering

Sustainable Materials & Byproduct Utilization

Life Cycle Assessment (LCA) and experimental characterization of industrial byproducts—ferrochrome slag, fly ash, olive cake ash, and zeolite—for eco-friendly concrete, high-temperature resilience, and green soil stabilization.

Ferrochrome Slag Fly Ash & Zeolite Life Cycle Assessment (LCA) Green Concrete

Infrastructure Economics & Construction Project Analytics

Game-theoretic Nash equilibria for contractor-subcontractor win-win payment term negotiations, portfolio cash flow optimization, risk sharing, and multi-objective project scheduling under budget constraints.

Game Theory Cash Flow Optimization Portfolio Scheduling JCEM (ASCE)

APPA — Automated Pavement Performance Assessment

Research platform · Ten analysis tools · One unified console connecting multi-sensor raw data to auditable condition ratings.

APPA console running Crack Assessment (DL) on the Crack002 sample: FPN EfficientNet-B4 on CUDA, coverage 1.99%, IoU 0.523 against the supplied ground-truth mask.
Actual output — Crack002 sample, FPN · EfficientNet-B4 on CUDA. Coverage 1.99 %, IoU 0.523 against the supplied ground-truth mask. IoU is unforgiving on hairline cracks: a one-pixel offset along a three-pixel-wide crack costs a third of the score.
APPA Console v3.4 [Research Edition]
Interactive walkthrough

Surface Distress Segmentation (FPN · EfficientNet-B4 / SAM 3)

Model: FPN · EfficientNet-B4 Coverage: 1.99% Resolution: 2048×1536 Ground Truth IoU: 0.523

Animated illustration of the segmentation step; the figures above are from the real Crack002 run shown at the top of this section.

Ground-Penetrating Radar (GPR) Subsurface Layering

Frequency: 1.5 GHz Horn Antenna
Asphalt Surface Thickness: 115.4 mm (±1.2 mm)
Granular Base Course: 240.8 mm
Dielectric Permittivity (εr): 6.12

3D Line Laser Scanning & Rutting Profile

Accuracy: ASTM E1703 Standard
Left Wheel Path Rut Depth: 14.2 mm (Moderate)
Right Wheel Path Rut Depth: 8.7 mm (Minor)

Zero-Shot MLLM Synthetic Expert Assessment

Results in Engineering (2026)
System Query: Evaluate distress imagery #HK-PV-8802 and determine condition rating index.
Synthetic Expert (Multimodal LLM):

Visual Analysis: High-density fatigue (alligator) cracking detected across 18.4% of wheel path area. GPR dielectric anomaly indicates sub-base moisture intrusion at depth 120mm.

Recommended Deduction: 38.5 points.

Condition Rating: Poor (PCI = 61.5 / 100).

Verified against 5-expert consensus panel (Mean absolute error: 1.8%).

Interactive Fuzzy AHP & Deductive Index Simulator

Auditable Calculation
Deductive PCI: 68.5 / 100
Fuzzy AHP Health Index: Good (0.712 Membership)
Action Required: Preventive Maintenance (Overlay & Sealing)

Selected Publications

Selected peer-reviewed articles from 48 total publications (35 journal articles, 21 Q1, 26 first/corresponding author, >900 citations, h-index 17).

Automation in Construction (2025) Q1 Top Journal (Ranked 2/193)

Pavement thickness evaluation using GPR and fuzzy logic

Ali Fares, Tarek Zayed

Automation in Construction, Volume 175, 106236. DOI: 10.1016/j.autcon.2025.106236

Develops an automated GPR signal processing and fuzzy logic model for non-destructive layer thickness assessment across asphalt and granular base courses with sub-centimeter accuracy.

Results in Engineering (2026) Q1 Journal (IF 9.4, Ranked 5/178)

Leveraging large language models as synthetic experts: a methodology and validation framework for pavement performance assessment

Ali Fares, Tarek Zayed

Results in Engineering (2026). High-impact open-access research.

Presents a pioneering zero-shot framework deploying Multimodal LLMs (MLLMs) as synthetic experts to interpret distress imagery, validated against expert engineering panels.

Automation in Construction (2024) Q1 Top Journal

Automated rebar recognition and corrosion assessment of concrete bridge decks using ground penetrating radar

Ali Fares, Tarek Zayed

Automation in Construction (2024).

Introduces deep learning and GPR B-scan amplitude attenuation algorithms for automatic rebar localization, cover depth measurement, and corrosion mapping in reinforced concrete bridge decks.

Environmental Sustainable Soil Stabilization / Green Materials Sustainability & LCA

Life cycle assessment of fly ash, olive cake ash, and zeolite for environmentally sustainable soil stabilization

Ali Fares, et al.

Environmental Science and Sustainable Infrastructure Research.

Quantifies carbon reduction and environmental impacts using Life Cycle Assessment (LCA) for industrial byproducts (fly ash, olive cake ash, zeolite) in sustainable soil stabilization.

Green Concrete & Industrial Byproducts Materials Research

Characteristics of ferrochrome slag aggregate and its uses as a green material in concrete – A review

Ali Fares, et al.

Construction & Building Materials Science.

Comprehensive review of ferrochrome slag as coarse and fine aggregates in eco-friendly concrete, examining thermal endurance, strength properties, and high-temperature performance.

Journal of Construction Engineering and Management (2025) Q1 Journal (ASCE)

Using game theory to negotiate win-win payment terms between contractors and subcontractors

Ali Fares, Ashraf El-Sawy, Tarek Zayed

Journal of Construction Engineering and Management, 151(5).

Mathematical game theory model establishing cooperative Nash equilibria for contract payment terms, cash flow optimization, and risk sharing in major infrastructure projects.

Teaching, Mentorship & Grants

Teaching & Pedagogy

  • Delivered tutorials & labs across undergraduate/postgraduate courses at PolyU & SQU.
  • Certified under PolyU’s Becoming an Effective Teaching Assistant (BETA) program.
  • Specialized in Civil Engineering Computation, BIM, Infrastructure Management, and Data Science.

Student Supervision

  • Co-supervised research students to successful completion and peer-reviewed Q1 journal publication.
  • Mentored undergraduate honours and MSc thesis projects in deep learning & GPR signal analysis.

Grants & Industry Liaison

  • Technical contributor & writer for an awarded Innovation and Technology Fund (ITF) government grant.
  • 2-year research liaison with government infrastructure departments (2022–2024).

Let's Collaborate on Intelligent Infrastructure

I am open to academic faculty appointments, research collaborations, grant proposals, and industry partnerships in AI, Digital Twins, Sustainable Materials, and infrastructure asset management.

ali-i.fares@connect.polyu.hk
Department of Construction Management and Innovation, PolyU, Hong Kong