Machine learning research with positive real-world impact.

We develop and deploy industry-leading machine learning systems. Our initiatives have the power to uplift large populations, while advancing the field of artificial intelligence.

View all posts

Ambitious applied research, positive outcomes

Layer 6 unites research, engineering, and product teams to quickly translate theory into impactful real-world applications.

Research Highlights

Our research is supported by access to massive datasets, close collaboration with world renowned academic faculty, and is deployed in impactful applications.

Our research areas include:

  • deep learning and generative AI
  • model explainability and trustworthy AI
  • time series modelling
  • natural language processing

View all publications

Empirical Methods in Natural Language Processing

EMNLP 2026 | A Gradient Perspective on RLVR Stability and Winner-Advantage Policy Optimization

Abstract

Reinforcement learning with verifiable rewards (RLVR) improves language-model reasoning, but GRPO-style optimization remains prone to collapse. We analyse this instability through token-level gradient dynamics, deriving a taxonomy that predicts how updates affect next-token probabilities and entropy. The taxonomy shows that stability depends jointly on the advantage sign and token distribution under the current policy. Motivated by this finding, we propose Winner Advantage Policy Optimization (WAPO), a simple online clipped policy-gradient objective that updates only on positive-advantage completions. Across mathematical reasoning and multi-hop QA benchmarks, WAPO improves training stability and matches or outperforms baselines across multiple model families.

Empirical Methods in Natural Language Processing

EMNLP 2026 | Chartographer: Counterfactual Chart Generation for Evaluating Vision-Language Models

Abstract

Chart question-answering (QA) benchmarks aim to pose questions that require visual reasoning to correctly answer, but models can often reach solutions through shortcuts or prior familiarity with a chart based on their own background knowledge. To strictly evaluate visual reasoning, we propose counterfactual charts where the chart-question task remains fixed, but underlying chart and the corresponding answer are varied. We introduce Chartographer, a framework to reverse engineer charts into executable code, validate reconstruction fidelity, generate seed-controlled counterfactual variants, and derive new answers from executable QA logic. We apply this framework to existing chart QA datasets and evaluate proprietary and open-source vision-language models (VLMs), measuring variation sensitivity and generalizability. Counterfactual charts reveal failures hidden by single-chart performance: VLMs often fail to generalize after answering the original chart correctly. We find failures are most prevalent when updated charts require novel visual reasoning pathways.

Findings of Empirical Methods in Natural Language Processing

Findings of EMNLP 2026 | Unifying Conformal Language Tasks with In-Context Ensembles

Abstract

Many NLP tasks, such as summarization and extractive question answering, reduce to retrieving relevant content from documents under two constraints: coverage, retaining enough pertinent information to achieve some goal, and conciseness, removing as much irrelevant information as possible. Conformal prediction methods have been used to guarantee coverage, and must be optimized for conciseness through design of a score function. State-of-the-art scoring functions use hand-engineered LLM prompts asking the model to rate the importance of content, but manual prompt engineering is labor-intensive and task-specific. We introduce the Conformal Relevance framework which uses in-context learning example curation and ensembling to create a score function which maintains coverage while improving conciseness with minimal manual input, and demonstrate its application on seven NLP tasks. We also theoretically study the impact of diversity for ensembled conformal scores, giving a complementarity condition that characterizes when ensembling improves worst-case sentence scores, and a saturation bound on ensemble improvement.

Conference on Language Modeling

COLM 2026 | LakeQuest: A Three-Domain Benchmark for Grounded Question Answering across Data Lakes

Abstract

While modern question answering (QA) systems excel on clean, schema-aligned corpora, real-world knowledge is rarely so neatly packaged. Answering questions over enterprise and scientific data lakes requires systems to navigate heterogeneous, weakly structured collections of tables, passages, and linked metadata. Current benchmarks abstract away this noisy discovery process, failing to evaluate end-to-end performance. To bridge this gap, we introduce LakeQuest, a human-validated benchmark of 9,846 QA pairs designed to evaluate the end-to-end retrieve-and-synthesize pipeline over realistic data lakes. LakeQuest spans three diverse domains (AI/ML metadata, retail banking, and multimodal biomedical drug information) and pairs every question with exact, modality-aware evidence pointers. By isolating source discovery from cross-modal synthesis, LakeQuest exposes critical failure modes in modern QA systems. Our baseline evaluations, including standard Retrieval-Augmented Generation (RAG) and agentic tool-use methods, reveal that high-quality retrieval does not guarantee correct reasoning. Systems consistently struggle with relation chaining in metadata graphs, policy grounding in bank ledgers, and joint tabular QA in biomedical contexts, highlighting the need for robust discovery and faithful cross-file composition mechanisms in future agentic QA systems.

International Conference on Machine Learning

ICML 2026 | Agentic Monte Carlo: Simulating Reinforcement Learning for Black-Box LLM Agents

Abstract

LLM agents operate in two distinct regimes: open-weight agents amenable to reinforcement learning (RL) and black-box agents whose behaviour must be controlled purely at test time. Although black-box agents are often backed by state-of-the-art proprietary LLMs, API-only access precludes parameter-level optimization, rendering most RL methods inapplicable. To address this limitation, we turn to a known equivalence between RL and Bayesian inference. We propose Agentic Monte Carlo (AMC) to directly sample from the optimal policy of a black-box agent rather than training it through RL. The optimal policy is a posterior over trajectories whose prior we define as the fixed black-box LLM agent. We employ Sequential Monte Carlo to sample from this posterior by learning a value function to steer the agent while leaving the underlying black-box model unchanged. We validate AMC on three diverse environments from the AgentGym benchmark, demonstrating significant improvements over prompting baselines and even outperforming Group Relative Policy Optimization (GRPO) as we scale the test-time compute of our method. AMC demonstrates the feasibility of performing principled RL-style optimization of black-box LLM agents.

Our research areas include:

  • deep learning and generative AI
  • model explainability and trustworthy AI
  • time series modelling
  • natural language processing

View all publications

Big vision, deep roots

  • The founders of Layer 6 are also the founders of the Vector Institute for Artificial Intelligence, and we maintain active research collaborations with Vector faculty.
  • Signal 1 spun out of the research collaboration between Layer 6 and St. Michael’s Hospital, and provides cutting-edge AI platforms for real time monitoring of patient outcomes.
  • Radical Ventures was launched by the founders of Layer 6 to incubate and support leading AI startups in Canada and abroad. With over a billion dollars raised, Radical has become one of the premier AI focused venture funds in the world.

Impactful partnerships

Originally founded in 2011, Layer 6 now forms the AI centre of excellence of TD Bank Group. Layer 6 impacts the lives of over 27 million customers, helping more people achieve their financial goals and needs through AI systems founded on the responsible use of AI.

Layer 6 embraces opportunities to contribute to the Canadian AI ecosystem. The founders of Layer 6 played pivotal roles in launching the Vector Institute, Radical Ventures, and Signal 1. Together these entities are integral in driving the Canadian AI innovation, from research to product incubation to scale-up. We continue to collaborate with leading academic institutions globally.

Passion to learn, driven to succeed

Our team comes from globally diverse backgrounds and we care deeply about fostering an inclusive culture. We learn from each other and win together. We are united by our passion for deep learning and a desire to apply our skills to have an outsized and positive impact on the future.

Meet some of our team

  • Maksims Volkovs

    Co-founder, Layer 6
    SVP, Chief AI Scientist, TD Bank Group

    Maks is a co-founder of Layer 6 and the Chief AI Scientist at TD Bank Group. Maks received his PhD from the University of Toronto where he was part of the Machine Learning Group led by Geoffrey Hinton. He previously worked at Microsoft Research and at Credit Suisse in London, UK. Maks is the winner of multiple international machine learning competitions, including a 3x winner of the ACM RecSys Challenge, a leading international competition on Recommender Systems. He has published over 50 papers and is a co-inventor of over 200 patents.

    LinkedIn Google Research

  • Sumee Seetharaman

    VP, AI Centre of Excellence & Practice, Layer 6

    Sumee is the Vice President of the AI Centre of Excellence & Practice, Layer 6. In the last 12 years, Sumee has held various product leadership roles across Digital (Open Banking, Marketplaces, Digital Identity) and AI (Layer 6, Enterprise Innovation) at TD. Prior to TD, Sumee worked at large technology companies and has shipped enterprise, developer, and customer facing software products. She holds an MBA from the Rotman School of Management at the University of Toronto and a Masters in Electrical Engineering from the University of Cincinnati.

    LinkedIn

  • Guangwei Yu

    Associate Vice President, AI Engineering

    Guang is a Staff Machine Learning Scientist at Layer 6 and Associate Vice President of TD Bank Group. Guang leads engineering teams building machine learning engines to enable AI models at the bank, and research teams in the areas of computer vision, tabular data and optimization. He is a repeat winner in the ACM RecSys Challenge and medalist in multiple international machine learning competitions including back-to-back first place finishes in ICCV challenges. Guang received his MScAC in computer science and BASc in Engineering Science from the University of Toronto.

    LinkedIn Google Research

  • Jesse Cresswell

    Staff Machine Learning Scientist, Trustworthy AI

    Jesse is a Staff Machine Learning Scientist at Layer 6 and TD where he leads research in Trustworthy AI – a field that aims to make machine learning models more explainable, fair, robust, and private. He has co-authored more than 30 research papers on a wide range of topics including deep generative models, federated learning, and representation learning, and is a co-inventor on over 50 patents. In 2019 he completed a PhD in Physics at the University of Toronto where he was a Vanier Scholar.

    LinkedIn Google Research

  • Chad Koziel

    Associate Vice President, GenAI Delivery, Layer 6

    Chad leads AI product delivery at Layer 6. He received his bachelor’s from British Columbia Institute of Technology, and his master’s from Queen’s University. He joined the data science field in 2014, after early career stops in metallurgy, professional poker, supply chain and deal-making. He is now a published author, co-inventor of 5 patents (and growing) and has brought numerous AI products to market, including TD’s first generative AI chatbot. Chad is a husband, and father to two young children.

    LinkedIn

  • Wankun Li

    Associate Vice President, Product Delivery, Layer 6

    Wankun is an Associate Vice President leading AI product delivery and helping to grow and strengthen Layer 6’s presence across the U.S. market. Wankun holds a PhD from the University of California, San Diego, and brings deep applied AI expertise alongside extensive management consulting experience supporting global and Canadian banks in risk transformation, credit risk modeling, and regulatory readiness. Since joining Layer 6 in 2025, she has served as a product owner, delivering enterprise scale AI solutions and translating advanced capabilities into business ready outcomes.

    LinkedIn

  • Keyvan Golestan

    Sr. Machine Learning Scientist

    Keyvan is a Senior Machine Learning Scientist at Layer 6. He earned his PhD in Computer Engineering from the University of Waterloo, where he conducted research on probabilistic graphical models and their applications to connected cars at the Centre for Pattern Analysis and Machine Intelligence. Keyvan completed his Master’s degree in Machine Intelligence and Robotics from the University of Tehran, Iran, where his background in robotics contributed to his success in prior research roles focused on self-driving cars. Currently, his research focuses on generative AI for tabular data and meta/in-context learning.

    LinkedIn Google Research

  • Anthony Caterini

    Sr. Machine Learning Scientist

    Anthony is a Senior Machine Learning Scientist, leading a team responsible for delivering models in the finance & trading space. Prior to joining Layer 6, Anthony completed a PhD in the Department of Statistics at the University of Oxford, with a focus on statistical machine learning and generative modelling. He also completed a BMath and MMath at the University of Waterloo, with several internships focused on finance and research. Besides the applied side, Anthony has also helped deliver over ten research papers to top conferences and journals whilst at Layer 6, focusing on the areas of generative modelling, tabular data analysis, and anomaly detection.

    LinkedIn Google Research

  • Satya Krishna Gorti

    Staff Machine Learning Scientist

    Satya is a Staff Machine Learning Scientist at Layer 6, where he works on vision and language understanding research and leads the ML Architecture team that designs and builds large-scale ML systems. Satya enjoys coding and machine learning challenges and has won multiple medals in international machine learning competitions. He received his MScAC degree from the University of Toronto and worked at Uber’s self-driving division and Morgan Stanley prior to joining Layer 6.

    LinkedIn Google Research

  • George Stein

    Sr. Machine Learning Scientist

    George is a Senior Machine Learning Scientist leading a research team building predictive foundation models. He holds a PhD in computational astrophysics from the University of Toronto and completed his post-doctoral research at UC Berkeley & Lawrence Berkeley National Laboratory, focusing on developing ML techniques for large scientific datasets. He has co-authored more than 30 research papers on a wide range of topics including representation learning, deep generative modelling, and tabular data analysis.

    LinkedIn Google Research

  • Himanshu Rai

    Sr. Machine Learning Scientist

    Himanshu is a Senior Machine Learning Scientist at Layer 6 where he manages the Computer Vision team. He is a Kaggle competitions Master with multiple gold models in competitions such as the ACM RecSys Challenge and the Google Open Images Challenge. He graduated from the MScAC program at the University of Toronto and holds an undergraduate degree in electrical engineering from IIT Patna. Before moving to Toronto, he worked as a software developer at Cisco Systems India.

    LinkedIn Google Research

Develop your career at Layer 6

Why are we called “Layer 6”? Deep learning is inspired by the human brain. The cortex has six layers of neurons, where the sixth layer integrates the deeper processing to produce an action.

We’re growing our team with people driven to be at the cutting edge of machine learning in research, engineering, and impactful applications.

View current openings