About

Hossein Shirali

I'm an applied AI scientist and engineer building production-ready computer vision and machine learning systems for robotics, automation, and scientific imaging. I enjoy turning messy, real-world data into robust, testable, and deployable AI solutions.

Most of my work sits at the intersection of computer vision, robotics, and large-scale data processing. I have built AI components for automated microscopes, robotic specimen handling systems, predictive maintenance in the energy sector, and satellite-based environmental monitoring.

While I am currently a PhD researcher at Karlsruhe Institute of Technology (KIT), my mindset is product-driven: define the problem with stakeholders, design an end-to-end pipeline, measure impact, and ship.

I like working close to the product: partnering with engineers, domain experts, and business stakeholders to translate cutting-edge research into tools that are maintainable, user-friendly, and ready for production.

Technical Skills & Expertise

I specialize in the complete AI development lifecycle, from research and prototyping to production deployment. My toolkit covers modern deep learning, computer vision, and the infrastructure needed to run systems reliably at scale.

Core competencies:

  • End-to-end ML pipeline development: data engineering, model training, evaluation, and deployment into production environments
  • Computer vision & deep learning: convolutional networks, object detection, segmentation, and custom architectures (PyTorch, TensorFlow, Keras, YOLO, OpenCV)
  • Production-grade engineering: clean Python, modular codebases, REST APIs, testing, model optimization (ONNX), Git-based workflows
  • MLOps & infrastructure: Docker, Kubernetes, cloud deployment (AWS), HPC/SLURM for large-scale training, CI/CD automation
  • Robotics & hardware integration: real-time imaging systems, camera pipelines, hardware–software integration, embedded AI
  • User-facing AI tools: dashboards and internal tools (Streamlit, Gradio), scientific computing tools, and open-source utilities

I am comfortable owning an AI project from idea to running system, and enjoy collaborating with product, engineering, and research teams along the way.

Professional Experience

Doctoral Researcher – Applied AI

Karlsruhe Institute of Technology (KIT)

Karlsruhe, Germany Mar 2023 – Present

  • Lead AI development for the Entomoscope photomicroscope and DiversityScanner robotic ecosystem, enabling high-throughput automated insect digitization and analysis.
  • Design and implement end-to-end deep learning pipelines for species identification, sex classification, orientation detection, and anatomical segmentation from robotic imaging systems.
  • Develop InsectMorphoAI, an image-based software suite for automated estimation of insect length, volume, and biomass from 2D images.
  • Contribute to the EU Horizon Europe project FORSAID (Forest Surveillance with Artificial Intelligence and Digital Technologies) by building deep learning classifiers for regulated forest pest detection.
  • Collaborate with interdisciplinary teams of engineers, taxonomists, and ecologists to integrate AI workflows into production-ready robotic devices deployed in biodiversity monitoring and forest health surveillance.

Data Scientist Consultant

Modis (Baker Hughes)

Baker Hughes (via Modis) Feb 2022 – Mar 2023

  • Developed optimization algorithms for turbine maintenance activities, improving planning efficiency and profitability.
  • Led cloud data migration initiatives, improving data security, accessibility, and reliability across enterprise infrastructure.
  • Designed and implemented solutions using citizen development tools (Microsoft Power Platform) to streamline internal workflows.
  • Delivered training on O365 and data tools, increasing digital adoption and productivity across teams.
  • Built automated reporting pipelines that reduced manual work and improved the quality and timeliness of insights.

Junior Data Scientist

Plasive Technologies

Bologna, Italy Mar 2021 – Sep 2021

  • Developed CNN-based semantic segmentation models for satellite imagery to support environmental monitoring and carbon stock estimation.
  • Built multidimensional datasets for satellite images and designed remote sensing methodologies for biomass estimation.
  • Used QGIS and related tooling for geospatial analysis, image processing, and visualization.

Machine Learning Engineer (Intern)

Kiwitron

Bologna, Italy Sep 2020 – Jan 2021

  • Designed a semi-automatic image labeling approach using 3D LiDAR point clouds as part of a master’s thesis project.
  • Reduced manual annotation time for industrial computer vision applications while maintaining label quality.
  • Delivered a solution that fed directly into model development workflows for safety-related industrial AI systems.

Key Projects

Entomoscope 2.0 – AI-powered photomicroscope

Karlsruhe Institute of Technology (KIT) 2024 – Present

Lead developer for the AI components of Entomoscope 2.0, a comprehensively redesigned, low-cost, open-source photomicroscope for automated insect digitization.

  • Built the ENIMAS 2.0 software suite covering image capture, cropping, background standardization, morphometric analysis, and rapid taxonomic screening.
  • Designed a modular, plugin-style architecture with human-in-the-loop validation for expert review.
  • Delivered one-click installers and FAIR data export to integrate with collections and citizen science workflows.

FORSAID – AI for forest pest surveillance (Horizon Europe)

Horizon Europe FORSAID, KIT 2024 – Present

Contributing to the Horizon Europe project FORSAID (Forest Surveillance with Artificial Intelligence and Digital Technologies), developing image-based early warning systems for regulated forest pests.

  • Design datasets and deep learning models for pest detection and classification from trap images.
  • Adapt Entomoscope and InsectMorphoAI tooling for forest health monitoring workflows.
  • Support plant health services and regulators with scalable AI-assisted screening pipelines.

InsectMorphoAI – automated morphometric estimation

KIT · Open-source 2023 – 2025

First author and main developer of InsectMorphoAI, an open-source software suite for measuring insect length, volume, and biomass from 2D images.

  • Implemented both a Streamlit web app and command-line interface for batch processing.
  • Added Docker support for reproducible deployments across lab and cloud environments.
  • Integrated with imaging devices to convert specimen images into morphometric data at scale.

AI-powered robotic specimen handling

KIT · Robotics & automation 2023 – 2025

Developed computer vision and AI components for robotic systems that pick, position, and image invertebrate specimens using flexible grippers and automated imaging setups.

  • Created detection and classification models for small, ethanol-preserved specimens under challenging imaging conditions.
  • Worked with robotics engineers to ensure real-time performance and robust handling.
  • Maximized throughput to support downstream analysis and large-scale biodiversity workflows.

Industrial turbine maintenance optimization

Modis / Baker Hughes 2022 – 2023

Designed predictive maintenance and optimization workflows for gas turbines, using operational and sensor data to support scheduling and service decisions.

  • Developed machine learning models and optimization logic for maintenance planning.
  • Helped move from manual, static planning to data-informed decision-making.
  • Integrated results into reporting and internal tools for engineers and managers.

Carbon stock estimation from satellite imagery

Plasive Technologies 2021

Developed CNN-based semantic segmentation pipelines for remote sensing, enabling scalable estimation of carbon stocks and biomass from satellite data.

  • Constructed multi-dimensional datasets from satellite imagery and ancillary data.
  • Combined domain knowledge and ML to deliver actionable environmental insights.

Research & Publications

My academic work focuses on computer vision and automation for biodiversity, environmental monitoring, and industrial applications. For a complete list, visit my Google Scholar or ResearchGate profiles.

Peer-reviewed journal articles

  • Image-based recognition of parasitoid wasps using advanced neural networks.
    Shirali, H., Hübner, J. J., Both, R., Raupach, M. J., Reischl, M., Schmidt, S., & Pylatiuk, C. (2024). Invertebrate Systematics, 38(6), IS24011. DOI: 10.1071/IS24011
  • Automated handling of biological objects with a flexible gripper for biodiversity research.
    Wöhrl, L., Keller, L., Klug, N., Shirali, H., Meier, R., & Pylatiuk, C. (2024). Automatisierungstechnik, 72(7), 672-678. DOI: 10.1515/auto-2023-0238

Preprints & manuscripts

  • Automated specimen triage for dark taxa: Deep learning enables orientation, sex identification, and anatomical segmentation from robotic imaging.
    Shirali, H., Wuehrl, L., Lee, L., Klug, N., Meier, R., Pylatiuk, C., & Hartop, E. A. (2025). bioRxiv. DOI: 10.1101/2025.10.02.680063
  • InsectMorphoAI: Deep learning-based methods for automated estimation of insect length, volume, and biomass.
    Shirali, H., Ascenzi, A., Wuehrl, L., Beyer, N., Di Lorenzo, N., Vaccarella, E., Klug, N., Meier, R., Cerretti, P., & Pylatiuk, C. (2025). bioRxiv. DOI: 10.1101/2025.05.22.655251
  • Image-based recognition using advanced neural networks can aid surveillance of Agrilus (Coleoptera, Buprestidae) jewel beetles.
    Caruso, V., Shirali, H., Bouget, C., Curletti, G., de Groot, M., Groznik, E., Gutowski, J. M., Pylatiuk, C., Roques, A., Sallé, A., Sweeney, J., Wuehrl, L., & Rassati, D. (2025). ARPHA Preprints. DOI: 10.3897/arphapreprints.e154842
  • Classification of microplastic particles in water using polarized light scattering and machine learning methods.
    Saur, L., von Pawlowski, M., Gengenbach, U., Sieber, I., Shirali, H., Wuehrl, L., Kiko, R., & Pylatiuk, C. (2025). arXiv preprint. DOI: 10.48550/arXiv.2511.06901
  • Automated photogrammetric close-range imaging system for small invertebrates using acoustic levitation.
    Klug, N., Kramer, M., Mazrek, F., Wuehrl, L., Shirali, H., Meier, R., & Pylatiuk, C. (2024). TechRxiv preprint. DOI: 10.36227/techrxiv.172651022.21831566

In preparation

  • Entomoscope 2.0: A low-cost, AI-powered, open-source photomicroscope for automated insect digitization and morphometric analysis.
    Shirali, H., Wuehrl, L., Klug, N., Meier, R., & Pylatiuk, C. (in preparation). Manuscript in preparation.

Conferences & Talks

  • Shirali, H., Wuhrl, L., & Klug, N. (2025, March 18). AI-driven advances in Species Identification and Biomass Analysis [Conference presentation]. Entomology Congress (DGaaE 2025), Geisenheim, Germany. https://publikationen.bibliothek.kit.edu/1000180291
  • Shirali, H., Wührl, L., & Klug, N. (2024, August 26). AI as a Catalyst in Entomological Research by Simplifying Species Identification [Conference presentation]. 27th International Congress of Entomology (ICE 2024), Kyoto, Japan. https://publikationen.bibliothek.kit.edu/1000173979
  • Shirali, H., Wührl, L., Klug, N., Meier, R., & Pylatiuk, C. (2024, June 12). Advancing Biodiversity Research with AI-Driven Automation [Poster presentation]. Helmholtz Artificial Intelligence Conference (Helmholtz AI 2024), Düsseldorf, Germany. https://publikationen.bibliothek.kit.edu/1000171882
  • Shirali, H., Wührl, L., Klug, N., Meier, R., & Pylatiuk, C. (2023). Automated Biodiversity Research [Poster presentation]. Helmholtz Imaging Conference (2023), Hamburg, Germany. https://publikationen.bibliothek.kit.edu/1000159668

Education & Certifications

  • Ph.D. Candidate in Computer Science (2023–Present)
    Karlsruhe Institute of Technology (KIT), Germany
    Institute for Automation and Applied Informatics (IAI)
    Research Area Automated Image and Data Analysis (AIDA)
    Group Biomedical Engineering & Robotics (BER)
    Focus: deep learning for automated biodiversity research on invertebrates
  • Master of Engineering (M.Eng.) in Electronic Technologies for Big Data and IoT (2018–2021)
    University of Bologna, Italy
    Grade: 110/110 (summa cum laude)
    Thesis: “Semi-automatic image labeling method based on point clouds”
  • Bachelor of Engineering (B.E.) in Electrical and Electronic Engineering (2012–2017)
    Shahid Chamran University of Ahvaz, Iran

Professional certifications & self-directed learning

  • Machine Learning Specialization (Coursera, 2022)
  • Computer Vision (Kaggle, 2022)
  • Data Management and Programming (Kaggle, 2021): SQL, Python, machine learning
  • MATLAB Onramp (MathWorks, 2022)
  • IELTS English Proficiency: Score 6.5 (C1 level)

Contact

I'm actively interested in applied science and AI roles where I can design, build, and deploy real-world systems. If you are working on computer vision, robotics, or data-intensive products and need someone who can bridge research and production, feel free to reach out.

Download CV