We are currently finalizing speakers for SEMLA 2026. We have already lined up speakers from leading universities, research labs, and industry teams.
Stay tuned for more.
Invited Speakers
Speaker Bios
Peter Rigby is a Full Professor of Software Engineering at Concordia University in Montréal and a Software Engineering Researcher at Meta. His research studies how software developers collaborate to build successful systems, with a focus on code review, coordination, empirical software engineering, and the social structure of software teams.
His work is driven by evidence from real development settings and aims to identify practices that make software engineering more reliable, scalable, and effective. For SEMLA 2026, Peter brings deep expertise in the human and organizational side of software engineering, especially as agentic systems change how teams review code, coordinate work, and decide which parts of engineering should remain under human judgment.
Talk title: Moving Faster with Less Risk: DRS for Code Freeze and Code Review at Scale
Moving Faster with Less Risk: DRS for Code Freeze and Code Review at Scale
As software engineers increasingly direct, review, and integrate AI-assisted code, organizations need better ways to decide which changes are safe to ship and which deserve scarce expert attention. This talk presents a line of work on Diff/Defect Risk Score (DRS), a just-in-time defect prediction system that estimates whether a code change is likely to cause a production fault at Meta.
The talk begins with the deployment setting at Meta, where DRS has been remarkably successful. By using LLM-based risk prediction to distinguish safer changes from riskier ones, DRS enabled release processes to move away from broad code freezes toward more targeted, risk-aware deployment decisions. This reduced unnecessary blocking of low-risk work while still helping teams control production risk.
The same risk signal can also improve code review. Change safety depends not only on whether a diff is risky, but also on whether the right people review it. The Contribution-aware Changeset Safety Ratio captures the combined expertise of the author and the reviewer team, and motivates using risk signals to focus expert reviewers on the riskiest diffs. Building on this, the talk covers reviewer recommendation systems and live experiments that add or recommend expert reviewers for high-risk changes, increasing review depth and engagement with minimal impact on review latency.
Finally, the talk addresses an open problem: once the right reviewers are involved, where should they focus their attention? LLM-based DRS models can highlight risky hunks or files in a diff, and early results suggest these explanations can guide reviewers toward outage-causing regions. Turning model attention into trustworthy, actionable review guidance remains unsolved.
Together, these works show how defect prediction becomes most valuable when embedded into real engineering workflows: deciding what can safely ship, selecting the right reviewers for risky changes, and helping engineers focus attention where it matters most.
Related papers:
- Moving Faster and Reducing Risk: Using LLMs in Release Deployment
- Mitigating the Risk of Defects and Improving Knowledge Distribution with Code Reviewer Recommenders
- Code Reviewer Recommendation for High Risk Diffs at Scale: Workflow, Recommender, and Live Experiments
- A Preliminary Study on Explaining Risk of Code Changes using LLM-Based Prediction Models
- DRS-OSS: Practical Diff Risk Scoring with LLMs
Zhou Yang is an Assistant Professor in the Department of Computing Science at the University of Alberta and an Amii Fellow. His research focuses on trustworthy AI, automated software engineering, software engineering for AI, human-AI interaction, reinforcement learning, and large language models for code.
He completed his PhD at Singapore Management University, earned an MSc in Software Systems Engineering from University College London, and previously worked as a senior research engineer at the Centre for Research on Intelligent Software Engineering. His work examines how code models and AI-assisted development tools can be made more secure, private, efficient, and usable.
Talk title: Engineering Trusted Reinforcement Learning Agents
Engineering Trusted Reinforcement Learning Agents
Reinforcement learning has enabled the development of effective agents for critical tasks across domains such as robotic control, autonomous driving, AI coding assistance, and computer use. In this talk, I will discuss the trustworthiness challenges of RL-based agents. I will first uncover robustness issues in RL agents and show how curiosity-driven mechanisms can be leveraged to expose flaws in their decision-making. Next, I will discuss how to provide runtime verification and protection for these agents. We will also examine the hidden risks in the prevailing practice of fine-tuning agents on trajectories, including backdoor attacks and watermarking. Finally, I will share a broader outlook on the intersection of reinforcement learning and software engineering.
Dr. Nafiseh Kahani is an Assistant Professor in the Department of Systems and Computer Engineering at Carleton University. Her research focuses on software testing, AI-based systems testing, machine learning applications in software engineering, and software security. Her research lab is actively involved in industry-funded projects, including automated program and test repair, as well as data privacy in large language models. Her work has been published in top-tier venues, including the IEEE Transactions on Software Engineering (TSE).
Talk title: Testing Multi-Agent AI Systems
Testing Multi-Agent AI Systems
Multi-agent AI systems are increasingly used to coordinate agents, tools, permissions, and delegation across complex workflows. However, current evaluations often focus mainly on final task success, which may not show whether the system’s internal coordination structure has actually been evaluated. This talk introduces a new evaluation approach for multi-agent systems, focusing on how to evaluate agent reachability, tool access, permission boundaries, and delegation paths. The goal is to move beyond simple outcome-based evaluation and provide better evidence that multi-agent systems are evaluated in a meaningful and systematic way.
Dr. Yu Huang is an Assistant Professor of Computer Science at Vanderbilt University, where she directs the MIND Lab (Mixed INtelligence Development for Programming). Her research focuses on human-centered AI for software engineering, with an emphasis on computationally modeling human cognition and integrating those insights into AI systems for programming.
By combining methods such as medical imaging, eye tracking, and cognitive modeling, her work uncovers human cognitive processes to inform the next generation of AI-assisted programming tools. Her research also spans computer science education, human-AI collaboration, and the long-term sustainability of open-source software ecosystems.
Dr. Huang has received four ACM SIGSOFT Distinguished Paper Awards, the NSF CAREER Award, the 2025 ICPC Vaclav Rajlich Early Career Achievement Award, and recognition as a 2021 EECS Rising Star. Her work has been supported by NSF, GitHub, ARPA-H, and Vanderbilt University. She holds a PhD in Computer Science and Engineering from the University of Michigan.
Talk title: Decoding and Encoding Expertise: Toward Human-Centered AI for Software Engineering
Decoding and Encoding Expertise: Toward Human-Centered AI for Software Engineering
As AI becomes a fundamental partner in the software development lifecycle, a critical gap has emerged: while modern models excel at predicting code tokens, they remain cognitively blind to the high-level reasoning, intentionality, and memory constraints of human programmers. This cognitive misalignment results in AI models that often fail to match human judgment, leading to diminishing returns as task complexity increases. The first half of the talk focuses on decoding the developer. By establishing a systematic methodology to model developers’ cognitive processes, Dr. Huang uncovers the complex relationship between programming and natural language processing while providing a framework to quantify human expertise for software engineering. The second half focuses on encoding this expertise to drive measurable gains in AI for SE, moving from data-centric augmentation techniques to a cognitive theory-grounded approach. The results show that integrating human cognitive signals, such as expert attention patterns, significantly surpasses the performance of traditional cognitively-unaware AI models, with substantial improvements in tasks like code summarization and comprehension. This work establishes a closed loop between understanding human cognition and enhancing automated models for SE, providing a scalable foundation for the era of human-centered AI for software engineering.
Suhaib Mujahid is a Staff Machine Learning Engineer at Mozilla, where he leads projects at the intersection of software engineering and applied machine learning. His focus is on building AI agents, bridging the gap between cutting-edge research and practical engineering, and tackling complex software engineering challenges.
He received his PhD in Software Engineering from Concordia University in 2021, with research spanning mining software repositories, software ecosystems, release engineering, and machine learning on code. His work has been published at ICSE, FSE, ICSME, and MSR, as well as in TSE, EMSE, and TEM.
TBD
Qiaolin (Isabelle) Qin is a third year PhD student in the Department of Software Engineering at Polytechnique Montreal. Her research focuses on improving AI-based software trustworthiness and security.
Isabelle serves as a reviewer and organizing committee member at top journals and conferences. Her research interests include software monitoring, AI model explainability, and reverse engineering. At SEMLA 2026, she brings insights into the novel security issues brought by the integration of AI models, and how to address these challenges by enhancing software supply chain transparency.
TBD
Tushar Sharma is an Assistant Professor in the Faculty of Computer Science at Dalhousie University. His research focuses on software design and architecture, refactoring, code quality, technical debt, software maintenance, mining software repositories, and machine learning for software engineering.
Before academia, Tushar worked at Siemens Research in the United States and Siemens Corporate Technology in India. He has written extensively on software design smells, refactoring, and maintainability. His SEMLA fit is strongest around sustainable AI, ML for software engineering, maintainability, and the engineering discipline needed to build reliable AI-driven systems.
Talk title: Sustainability in AI Systems: Challenges and Opportunities
Sustainability in AI Systems: Challenges and Opportunities
Artificial Intelligence (AI) models are energy-hungry. As these models grow larger and become more complex, their energy consumption increases substantially. This leads to higher carbon emissions and rising operational costs, creating serious challenges for sustainable computing. This talk highlights the environmental impact of modern AI systems, introduces key metrics for measuring energy use and carbon footprint, and discusses available tools and frameworks for measuring and monitoring energy consumption. Furthermore, the session summarizes current research and practical techniques for developing greener AI systems, including model pruning, quantization, efficient architecture design, and workload scheduling, and situates these within the wider movement toward responsible and sustainable AI development. The talk concludes by highlighting important open research challenges and calling on the community to treat energy efficiency not as an afterthought but as a first-class objective in the design and evaluation of AI systems.
Nikita Dvornik is an AI researcher affiliated with RBC Borealis. His work is in machine learning and computer vision, with emphasis on representation learning, visual understanding, and models that can generalize from limited supervision.
His background in perception and learning systems adds a strong perspective on robustness, generalization, and model behavior under real-world constraints. For SEMLA 2026, his work connects AI research to high-stakes production settings where models need to perform reliably beyond controlled benchmarks.
Talk title: Foundation Models for Financial Services
Foundation Models for Financial Services
This talk explores the emerging paradigm of foundation models trained on transaction data: large-scale models that learn behavioural embeddings from billions of financial events, much like LLMs learn from text. We discuss how these models are built and used, on their own or in tandem with LLMs, and share how they unlock powerful downstream applications across credit decisioning, fraud detection, and personalized advisory — opening a new chapter in AI for financial services.
Lovedeep Gondara is Head of AI Research and Development at Vanguard and an Adjunct Professor at the University of British Columbia. He is a machine learning researcher whose work spans large and small language models, agentic methods, model grounding, privacy-preserving machine learning, differential privacy, deep learning, decentralized learning, statistics, and multimodal machine learning.
He earned his PhD in Computer Science from Simon Fraser University and has worked on applying machine learning in sensitive domains where reliability, privacy, and governance matter. At SEMLA 2026, Lovedeep brings a regulated-industry perspective on trustworthy AI, responsible deployment, and language-model systems that need to operate under real organizational constraints.
Talk title: The Importance of Trustworthy AI in Heavily Regulated Domains
The Importance of Trustworthy AI in Heavily Regulated Domains
The deployment of AI systems in heavily regulated domains such as healthcare and finance presents unique challenges that extend beyond technical performance metrics to encompass fundamental questions of trust, accountability, and societal impact. Both sectors share critical characteristics: they involve high-stakes decisions with material consequences for individuals, operate under stringent regulatory oversight, and are fundamentally client-facing, requiring that end-users place significant trust in system outputs that may influence their health outcomes or financial wellbeing.
This talk examines the core pillars of trustworthy AI in these contexts, including transparency and explainability of model decisions, robustness to distribution shift and adversarial inputs, fairness across demographic groups, rigorous validation against domain-specific standards, and mechanisms for human oversight and intervention.
Trustworthiness is a practical necessity, not merely an ethical desideratum. Systems that fail to earn stakeholder trust face regulatory barriers, adoption resistance, and reputational risk. By drawing parallels between both heavily regulated domains, the talk identifies transferable lessons for building AI systems that meet the elevated standards these domains demand.
Vincent Fortier leads Field Engineering for Databricks’ Public Sector business in Canada. His team helps governments, hospitals, and universities move AI and ML workloads from pilot to production, with the constraints of citizen data, PHI, and public accountability baked in.
With 20+ years on the technical side of enterprise data and machine learning, his focus is on what it actually takes to ship ML systems that hold up in the real world: trust, scale, compliance, and the engineering discipline that gets you there.
AI is being deployed faster than the controls around it, and security teams are the ones paying the bill. In this session, Vincent shows how Databricks gives you a single governed foundation for data, AI, and agents, so innovation doesn’t outrun oversight. The talk closes with a look at Lakewatch, Databricks’ open agentic SIEM built for machine-speed defense.
Patrice Béchard is an Applied Research Scientist at ServiceNow in Montréal, where he works on AI agents for enterprise automation. His research studies how large language models can reliably interact with complex enterprise systems to execute real-world workflows.
His recent work includes workflow generation, business-process automation, web agents, agent debugging, retrieval-augmented generation, and reducing hallucinations in structured outputs. Patrice fits SEMLA 2026 directly because his work sits at the center of enterprise agents: how they act, how they are evaluated, and how teams can make their outputs reliable enough for production use.
TBD
Gustavo Pinto is an AI Engineer at Zup Innovation and a software engineering researcher with experience across research and industry. His work spans open-source software, human aspects of software engineering, empirical software engineering, mining software repositories, AI agents, foundation models, and machine learning for software engineering.
He has published widely in software engineering venues and works on problems that connect developer productivity with intelligent software tools. At SEMLA 2026, Gustavo brings a practical view of how software teams adopt AI systems, how developer workflows change, and what evidence is needed before organizations trust these tools.
TBD
Orlando E. Marquez is a Lead Applied Research Scientist at ServiceNow with a strong background in software engineering. His research focuses on natural language processing, text-to-text systems, explainability, error analysis, semi-supervised learning, summarization, structured outputs, and text-to-workflow generation.
Orlando has worked on shipping NLP and GenAI systems to enterprise users, including low-code workflow generation and structured outputs from natural-language requirements. His SEMLA contribution is strongest around turning AI research into deployed enterprise systems with careful evaluation, engineering constraints, and user-facing reliability.
TBD
Zhen Ming (Jack) Jiang is an Associate Professor and York Research Chair at York University in Toronto. His research spans software engineering for and with AI, software performance engineering, and software maintenance.
He is a co-organizer of the Shonan Meeting on “Foundation Models and Software Engineering” and has co-organized AIWare events including FM+SE Summit 2024 and the 2025 AIWare Bootcamp. He received his PhD from Queen’s University and previously worked with the Performance Engineering team at BlackBerry, where tools from his research are used daily to monitor and debug ultra-large commercial software systems.
TBD
Dr. Tanja Tajmel is an Associate Professor at the Centre for Engineering in Society, Gina Cody School of Engineering and Computer Science, Concordia University. She earned her PhD in Didactics of Physics from Humboldt-University Berlin and holds a BSc in Meteorology and Geophysics and an MSc in Physics and Philosophy from the University of Graz, Austria.
Internationally recognized for her leadership in developing frameworks for socially equitable research and education in science and technology, she has led several inter- and transdisciplinary initiatives including a think tank exploring STEM education as a human right. Since 2021, she has been Co-PI in the pan-Canadian NSERC CREATE program SE4AI, where she designed the graduate course Social Aspects of AI. Since 2024, she serves as Scientific Director for Equity, Diversity and Inclusion at Volt-Age, Concordia’s large-scale research program supported by the Canada First Research Excellence Fund.
Talk title: Social Dimensions of AI: Challenges and Opportunities for Software Engineering Education
Social Dimensions of AI: Challenges and Opportunities for Software Engineering Education
Artificial intelligence has fundamentally transformed a wide range of industries over the past decade and is expected to reshape nearly every aspect of society in unprecedented ways. Software developers are at the forefront of this development, both as creators and as professionals whose working lives are shaped by the very technology they build. Rather than viewing AI as a neutral set of tools, it is becoming clear that it is embedded in complex social, economic, and ecological systems. From dependence on global networks of labor, data, and natural resources to the diverse applications of AI systems, these technologies raise critical questions about ethics, responsibility, and what we as humans consider a decent life. The talk explores how higher education in science and engineering fields, particularly in software engineering, can respond to these pressing questions of our times.
Boqi (Percy) Chen is an Assistant Professor in the School of Electrical Engineering and Computer Science at the University of Ottawa. His research focuses on the reliable and robust integration of AI components, particularly large language models, into software engineering processes.
He received his PhD from McGill University and during his doctoral studies worked as an R&D Engineer at Aggregate Intellect through a Mitacs collaboration and as a part-time Research Associate at the Huawei Waterloo Research Center. For SEMLA 2026, his work emphasizes improving the reliability of AI-based software engineering automation through approaches that provide provable guarantees.
Talk title: Reliable Generation of Software Artifacts with LLMs: From Ad-Hoc Validation to Semantics-Aware Generation
Reliable Generation of Software Artifacts with LLMs: From Ad-Hoc Validation to Semantics-Aware Generation
Software engineering demands artifacts that satisfy rigorous syntactic and semantic constraints, yet large language models are inherently probabilistic and non-deterministic. As LLMs are increasingly used to automate software engineering tasks, ensuring the reliability of the artifacts they produce becomes a critical challenge. This talk presents recent efforts toward reliable AI-based generation of software artifacts, particularly domain-specific languages and other structured representations. It first demonstrates how current LLMs frequently violate both syntactic and semantic constraints, then presents two complementary approaches: AbsCon, a post-processing framework that leverages self-consistency to ensure provable specification compliance, and Projectional Decoding, a semantics-guided decoding approach that proactively steers generation toward valid outputs. Together, these approaches illustrate emerging directions toward LLM-based software automation that developers can trust.