Ali Lesani

Ali Lesani

Ali Lesani

AI and Computer Vision Researcher

PhD Student at CViSS Lab

Hi, I'm Ali!

I am an AI researcher and PhD student at the University of Waterloo, working on vision–language models for spatio-temporal reasoning in real-world inspection and monitoring systems. My research focuses on interpretable visual understanding, including component-level change detection, 3D scene understanding, and multimodal reasoning. I have hands-on experience designing and deploying production-grade AI systems through collaborations with academic and industry partners. My expertise includes Python and modern ML frameworks (PyTorch, Hugging Face, scikit-learn, OpenCV), with a strong focus on building end-to-end machine learning systems from research to deployment.

Experience

ML Engineer

Kisoji Biotechnology

2025 - Present

AI & ML Researcher (PhD)

CVISS Lab, University of Waterloo

2023 - Present

View All Experience →

Education

PhD, Engineering

University of Waterloo

2023 - Present

M.Sc., Engineering

Sharif University of Technology

2019-2022

View All Education →

Technical Skills & Tools

Python

Python

C++

C++

PyTorch

PyTorch

HuggingFace

HuggingFace

LangChain

LangChain

scikit-learn

Scikit-learn

OpenCV

OpenCV

FastAPI

FastAPI

Streamlit

Streamlit

Pandas

Pandas

SQL

SQL

AWS

AWS

Docker

Docker

Nginx

Nginx

GitHub Actions

GitHub Actions

Git

Git

Bash

Bash

Linux

Linux

Experiences

Machine Learning EngineerAug 2025 - Present

Kisoji Biotechnology Inc.

  • Architected and deployed an agentic RAG platform enabling natural-language querying over internal documents and structured databases, implementing end-to-end RAG pipelines and intelligent tool routing between document retrieval and NL2SQL, with CLI and Streamlit interfaces deployed on AWS.
  • Built an automated PDF image extraction and classification system using embedding-based similarity against reference images, exposed via FastAPI REST API with web UI and CLI support.
  • Architected and deployed scalable cloud infrastructure on AWS, configuring VPCs, subnets, Internet Gateways, routing tables, security groups, and IAM roles; containerized multi-service applications using Docker and Docker Compose, served via Nginx reverse proxy, and automated deployments using GitHub Actions CI/CD.

NLP Specialist (RA)Jan 2025 - Aug 2025

Office of Associate Dean, Teaching & Student Experience, University of Waterloo

  • Contributed to the development of an automated NLP-based comment classification system for student course evaluations, using Sentence Transformer (all-MiniLM-L6-v2) embeddings and a KNN classifier to identify and filter inappropriate feedback.
  • Implemented Python pipelines for data preparation, model training, evaluation, and batch inference.

AI & Machine Learning Researcher (PhD Student) Sep 2023 - Present

CViSS Lab, University of Waterloo

  • PhD Research
    • Developing a spatio-temporal vision–language system for automated infrastructure inspection, enabling component-level change detection across multi-cycle UAV imagery by reasoning over time with text-conditioned VLMs that produce interpretable natural-language change reports and grounded pixel-level evidence.
  • Mitacs Intern | Nav CanadaJan 2026 - Present
    • Developed a computer vision system to compute Runway Occupancy Time (ROT) by detecting and tracking aircraft in airport video feeds, implementing small-object detection using SAHI combined with object detectors and monitoring runway-entry and exit events to accurately measure aircraft presence during takeoff and landing.
  • Mitacs Intern | RBC Royal BankJan 2025 - Apr 2025
    • Developed a monocular window measurement system for energy auditors to estimate window dimensions from single images, integrating Grounded SAM for window segmentation, monocular metric depth estimation, and camera-geometry–based 3D projection using known intrinsics to compute accurate width and height. [ GitHub]
  • Mitacs Intern | Rogers CommunicationsSep 2024 - Dec 2024
    • Developed SDG-SAM, a zero-shot instance segmentation pipeline for fine-grained detection of cell tower components in UAV imagery, integrating saliency-based foreground extraction, monocular depth estimation (Depth Anything), open-vocabulary detection (Grounding DINO), and foundation-model segmentation (SAM) to handle complex outdoor scenes with heavy clutter and occlusions. [ GitHub]

Software Developer & Graduate Researcher Sep 2020 - May 2023

INSURER Lab, Sharif University of Technology

  • Worked as a developer on Rtx, a C++ and Qt-based simulation software for probabilistic modeling, reliability, and resilience analysis of large-scale infrastructure systems.
  • Implemented agent-based models to simulate disaster impacts and interdependent infrastructure behavior.
  • Developed modules for assessing community-scale risk and resilience across varying levels of model refinement.
  • Contributed to probabilistic simulation workflows for evaluating post-disaster recovery and system performance.

Selected Projects

Intelligent Decision-Making System: Deep Learning-Based Satellite Imagery Analysis for Probabilistic Inference via Bayesian Networks

  • Developed a deep learning pipeline for satellite imagery analysis, segmenting building footprints using U-Net and classifying structural damage using ResNet-based feature representations.
  • Integrated vision-based damage outputs into a Bayesian Network to enable probabilistic inference of community-level impact and uncertainty-aware decision support.
  • Designed a modular, updateable inference pipeline that incorporates new imagery or evidence to refine predictions over time.

Education

Doctor of Philosophy, Engineering 2023 - Present

University of Waterloo

GPA: 99.67/100 (4/4)

Master of Science, Engineering 2019-2022

Sharif University of Technology

GPA: 17.83/20 (3.87/4)

Bachelor of Science, Engineering 2015-2019

Sahand University of Technology

GPA: 19.12/20 (3.96/4)

Honors & Awards

1st Place, 2025 NSF NHERI GSC Data Challenge2025

Awarded for top AI research in natural hazards.

Winner of UW Graduate Scholarship2024

University of Waterloo.

Winner of the National Elites Foundation Award2019

Ranked 1st Among More than 44 Peer B.Sc. Students2019

Sahand University of Technology.

Ranked 49th out of 30,000+ in Nationwide M.Sc. Entrance Exam2019

Ranked 1st, 24th Regional Scientific Olympiad in Civil Engineering2019

Top Undergraduate Student in Department2016 and 2017

Recognized for academic excellence.

Member of the Center of Exceptional Talents2015 - 2019

Sahand University of Technology.

Publications

MonoWindow publication thumbnail

MonoWindow: Monocular 3D Window Measurement in Indoor Environments

Ali Lesani, Chul Min Yeum
In Preparation.
Knowledge-augmented GPT publication thumbnail

A Knowledge-Augmented GPT for Efficient Data Collection in Home Energy Audits

Shih-Chi Liu, Kay Han, Ali Lesani, Chul Min Yeum, Graham Watt, Seungjae Lee
Proceedings of the 6th International Conference on Building Energy and Environment (COBEE 2025).
Housing Recovery publication thumbnail

Simulating Housing Recovery Challenges Due to Post-disaster Repair Cost Increases

Ali Lesani, Alireza Ahmadi, Rodrigo Costa
Proceedings of the 18th World Conference on Earthquake Engineering (WCEE 2024).