Dr. Issam Laradji
Senior Staff Research Scientist, ServiceNow
Adjunct Professor, University of British Columbia
Founder and Director, Vancouver AI Meetup
Vancouver, British Columbia, Canada
Email | Google Scholar | LinkedIn | GitHub
Summary
Issam Laradji is a Senior Staff Research Scientist at ServiceNow and an Adjunct Professor at the University of British Columbia. He earned his Ph.D. in Computer Science from UBC in 2020 and completed his postdoctoral training in Computer Science at McGill University from 2020 to 2021.
His research covers evaluation and retrieval methods for AI agents, data analytics agents, deep-research systems, multimodal models, and computer vision, including benchmarks for enterprise deep research and business analytics, retrieval from long documents, and evaluations of agent memory and skills.
His technical areas include foresight-driven agents for time series, deep research, analytics, and computer use; natural language processing; computer vision and graphics; optimization; benchmarking; and reproducible experimentation.
He has 70+ publications from 2013 to 2026 and 5,700+ citations. They include papers published at ICLR, NeurIPS, CVPR, ECCV, ACL, EMNLP, AISTATS, TMLR, WACV, and Nature Scientific Reports. Current citation totals are available on Google Scholar.
Professional Experience
Senior Staff Research Scientist, ServiceNow
Feb 2024 - Present
- Lead the Predictive and Insights Agents team and publish research on enterprise deep-research benchmarks, multi-step business analytics agents, skill-adaptive insight discovery, retrieval, and agent memory, represented by DRBench at ICLR 2026, InsightBench at ICLR 2025, and AgentAda at ACL 2025.
Staff Research Scientist, ServiceNow
Aug 2021 - Feb 2024
- Published research on web agents and workflow discovery, LLM-based intent discovery and low-resource NLP, multimodal generation, few-shot and self-supervised vision, counterfactual explanations, continual learning, and optimization, including WorkArena at ICML 2024, IntentGPT at ICLR 2024, PromptMix at EMNLP 2023, and Kubric at CVPR 2022.
Adjunct Professor, University of British Columbia
Feb 2023 - Present
- Supervise student research in AI agents, retrieval, language models, and computer vision.
- Advisor of the UBC Master of Data Science Vancouver External Academic Advisory Board for 2026 to 2028.
Postdoctoral Researcher, McGill University
Aug 2020 - Aug 2021
- Published research under Derek Nowrouzezahrai on weakly supervised segmentation, image classification, 2D-to-3D reconstruction, counterfactual explanations, and adaptive stochastic optimization, including Affinity LCFCN in Scientific Reports and Beyond Trivial Counterfactual Explanations at ICCV 2021.
Research Intern
ServiceNow, Jan 2021 - Aug 2021
Element AI, Sep 2017 - Dec 2020
- Published research on few-shot learning, weakly supervised localization and segmentation, continual learning, domain adaptation, and stochastic optimization, including Embedding Propagation at ECCV 2020, Where Are the Blobs at ECCV 2018, and Painless Stochastic Gradient at NeurIPS 2019.
Selected Research Contributions
Deep-research and analytics evaluation
- DRBench, ICLR 2026: benchmark for enterprise deep-research tasks.
- InsightBench, ICLR 2025: evaluates business analytics agents through multi-step insight generation.
- DRFLOW, arXiv 2026: benchmark for predicting personalized deep-research workflows.
- Dr-CiK, arXiv 2026: testbed for evaluating foresight-driven agents.
Retrieval and generation
- SproutRAG, arXiv 2026: tree search with progressive embeddings for retrieval from long documents.
- MCompassRAG, arXiv 2026: uses topic metadata for paragraph-level retrieval.
- Dual-Confidence Contrastive Decoding, arXiv 2026: contrastive decoding method for retrieval-augmented generation.
Agent privacy, memory, and skills
- MosaicLeaks, arXiv 2026: studies privacy risks when deep-research agents query information in open environments.
- Are Online Skill and Memory Modules Always Worth Their Tokens?, arXiv 2026: evaluates web-agent skill and memory modules under token budgets.
Multimodal models and computer vision
- StarVector, CVPR 2025: generates scalable vector graphics code from images and text.
- AlignVLM, NeurIPS 2025: aligns vision and language latent spaces for multimodal understanding.
- BigDocs, ICLR 2025: dataset for document and code tasks in multimodal models.
- A Weakly Supervised Consistency-Based Learning Method for COVID-19 Segmentation in CT Images, WACV 2021.
Publication output
- Produced 20 research manuscripts in 2025, with 19 accepted and one submitted, spanning ICLR, NeurIPS, ACL, EMNLP, NAACL, CVPR, AISTATS, TMLR, workshops, demonstrations, and a U.S. patent. Projects included BigDocs, InsightBench, StarDoc, AlignVLM, LitLLM, AgentAda, EDRA, and DRBench.
- Published more than 50 peer-reviewed papers from 2013 to 2024 in machine learning, natural language processing, computer vision, and optimization. See the Google Scholar profile for the complete list.
Education
Postdoctoral Fellowship, Computer Science, McGill University
2020 - 2021, supervised by Derek Nowrouzezahrai
Ph.D., Computer Science, University of British Columbia
2014 - 2020, GPA: 4.0/4.0, supervised by Mark Schmidt
Thesis: Where Are the Objects? Weakly Supervised Methods for Counting, Localization and Segmentation
M.Sc., Information and Computer Science, KFUPM
2012 - 2014, GPA: 4.0/4.0
B.Sc., Information and Computer Science, KFUPM
2008 - 2012, GPA in final two years: 4.0/4.0
Academic Service
- Area Chair, NeurIPS 2026
- Area Chair, ACL Rolling Review, May and August 2026 cycles
- Advisor, UBC Master of Data Science Vancouver External Academic Advisory Board, 2026 - 2028
- Workshop organizer, CVPR Workshop on Learning from Limited Labelled Data for Image and Video Understanding, 2022; CVPR L3D-IVU Workshop, 2023 and 2024
- Challenge chair and lead, CVPR Continual Learning Workshop, 2020
- Reviewer, AISTATS, TPAMI, ICML, NeurIPS, CDMA, BMVC, ICLR, WACV, AAAI, MAIS, CVPR, ECCV, ICCV, EMNLP, Neural Networks, and Neurocomputing
Teaching and Community Programs
Founder and Director, Vancouver AI Meetup
Oct 2024 - Present
- Founded Vancouver AI Meetup in 2024 and grew it to more than 500 members. Lead weekly paper reading groups, biweekly AI-trends sessions and social events, workshops, hackathons, and project demonstrations.
- In 2025, ran four six-week app-building cohorts, four Python crash courses, a six-hour n8n automation workshop, and four Show and Tell events.
- Co-organized the Agent Arena 2.0 deep-research hackathon and Agentic AI 1.0 hackathon at Northeastern University Vancouver in 2025.
- Led the sixth Build Your AI Idea program from Jul 24 to Aug 23, 2026. The four-week cohort had 35 students and included weekly lectures, co-working sessions, social events, and app demonstrations.
- Ran a two-week personal AI assistant course in Apr 2026. Participants built assistants and presented them at a final demonstration.
- Completed the fifth Build Your AI Idea in Six Weeks cohort in Feb 2026, covering Flask, n8n, Lovable, React, and Node.js through lectures, labs, office hours, and co-working sessions.
- Co-hosted a beach BBQ for more than 100 attendees with Maple Leaf Venture Club for Web Summit Vancouver on May 10, 2026.
Additional teaching
- Taught VAM! Build an App with Claude Code in Under 3 Hours on Jun 14, 2026, guiding coders and non-coders through designing, building, and demonstrating their own local applications.
- Taught a full-day beginner course on building an AI app with Python on Jan 11, 2026. Students built and presented an app during the course.
- Co-facilitated n8n 101: Workflow Automation in Feb 2026. Participants built automated workflows, chatbots with tool use and memory, and daily-feed summarizers.
- Delivered an AI workshop for West Coast Family Centres on May 26, 2026. The session covered writing case notes, drafting email, finding relevant family programs, privacy, and bias.
- McGill University Capstone, large-scale experimentation frameworks, 2020 to 2021.
- UBC CPSC 340, Machine Learning and Data Mining, 2015 to 2016; CPSC 210, Software Construction, 2015; CPSC 221, Algorithms and Data Structures, 2015; CPSC 422, Advanced Artificial Intelligence, 2015; CPSC 322, Artificial Intelligence, 2014.
Research Software
- Created Haven-AI, a Python framework for launching, managing, and visualizing thousands of reproducible experiments used in research papers, competitions, and AI systems.
- Built AgentPoirot, an analytics agent that produces insights based on a user's goal and persona. Video demo
- Use Python and PyTorch for research software and have created or contributed to hundreds of open-source repositories. GitHub profile
Student and Intern Supervision
- 2026 research interns: Shambhavi Mishra, an École de technologie supérieure PhD student supervised by José Dolz and Marco Pedersoli; Raymond Li, a UBC Computer Science PhD student researching model interpretability and visualization; and Md Tawkat Islam Khondaker, a UBC Computer Science PhD student.
- Curtis Fox, UBC: DRBench; Amirhossein Abaskohi, UBC: 2026 retrieval papers; Amrutha Ramesh, UBC: NeurIPS workshop research.
- Gaurav Sahu, University of Waterloo: EMNLP research; Yuhongze Zhou, McGill: computer-vision research; Michael Noukhovitch, MILA: language-drift research.
- Led a McGill capstone team that built a mobile computer-vision application for recognizing objects in convenience stores.
Selected Talks
- AI in Finance Summer School, RBC Borealis, Aug 27, 2026: invited technical talk on agentic AI.
- West Vancouver Memorial Library, Jan 19, 2026: Tech Talk on Agentic Artificial Intelligence
- UBC Applied Mathematics Workshop, Nov 15, 2025: presentation on DRBench.
- Microsoft AgentCon, Nov 2025: hands-on tutorial on deep-research agents and benchmarks.
- AgentPoirot podcast, Aug 2025: discussion of agentic AI research and deep-investigation systems. Video
- BC Library Association, Oct 16, 2024: Best Practices for Hosting AI Workshops at the Public Library
- UBC, McGill, KFUPM, and YouTube, 2020 to 2024: tutorials and demonstrations on large-scale experimentation, language models, 2D and 3D computer vision, and optimization.
Miscellaneous
- Had an elevator pitch with Yoshua Bengio about my research at Element AI in 2020.
- Founded Board Games Social - Vancouver in 2023, a community for relaxed weekly and biweekly board-game gatherings and meeting new people.
Awards
- MITACS Accelerate Postdoctoral Award, 2020
- UBC Four Year Fellowship, 2016 - 2020
- UBC Teaching Assistant Award, 2015
- Big Data University Innovation Prize, SportsHack, 2015
- Google Summer of Code, 2014