NDC Toronto 2026
2026-05-05..2026-05-08
Toronto, Canada
as the organisers put it
NDC Toronto 2026 is a 4-Day Event for Software Developers, 5-8 May at The Westin Harbour Castle, Toronto.
who it's for
software developers, machine learning engineers, engineering managers, enterprise architects, platform engineers, security engineers, compiler developers
topics
deploying ML models to production; agentic system testbeds; LLM-powered developer workflows; prompt security and supply chain risk; Rust MCP gateway integration; Svelte server-side rendering; modular monolith migration with DDD; SQL MCP server; MRI diagnostic ML case study; Carbon compiler benchmarking and optimization; writing Windows kernel drivers
| title | author | length |
|---|---|---|
| Why Great Models Fail: Lessons From 9 Years of Deploying ML Models | Megan Robertson | 59m |
After nine years deploying ML models, the main claim is simple: high test metrics alone don’t get models into production. An NLP proof of concept builds topic… | ||
| Building a Doom-Like World to Explore Agentic Systems | Alexander Chernov | 53m |
Builds a controlled Doom-like testbed that mirrors game state into an agent layer so actions are observable attributable and reproducible; implements a… | ||
| Three Lessons from a Woman in Enterprise Architecture | Sarah Wimberley | 47m |
Career arc moves from art and photography through marketing into enterprise architecture, founding Olster Technologies and shifting a role from 10% EA work to… | ||
| Like Having an Intern? The Impact of LLMs on Software Engineering Careers | Tom Sellek, Netta Avnoon | 01h 01m |
A preliminary survey of software engineers reports about 90% adoption of LLM tools, 80% using them multiple times a week, and roughly 3/4 positive sentiment… | ||
| Techniques for Wielding Data and Behaviors | Mike Shah | 51m |
Shows concrete code (mostly D with some C++) that builds a mini VMlike command queue and several data/behavior abstractions so you can reproduce them: bit… | ||
| Incrementally Rebuilding Stack Overflow's Monolith | Connell Sharp | 54m |
Stack Overflow incrementally rewrites the comments UI inside a modular monolith, rebuilding the frontend in Svelte while keeping .NET Razor server rendering. A… | ||
| Focus in a Fragmented World: A Developer’s Guide to Staying Productive | Susanne Bach Ladefoged Hou | 01h 02m |
Claims practical, battle-tested tactics for developers with ADHD: use time blocking and ramp in and ramp out to protect flow, apply the traffic light energy… | ||
| Be curious, not judgemental | Sergiy Dybskiy | 51m |
Wins a Bolt hackathon and won $100,000 which pays off a parents’ mortgage, out of 128,000 participants and 9,575 projects; builds a meal-planning app called… | ||
| Becoming an Engineering Manager Without Losing Your Technical Soul | Paul Williams | 01h 01m |
Argues that an engineering manager can stay deeply technical by doing strategic troubleshooting and building lightweight observability: atomic accumulated… | ||
| Calm at Scale | Oriana Sparks | 38m |
Claims attention is the fastest‑burning resource: average focus lasts 47 seconds and post‑interruption recovery averages 23 minutes. Offers concrete practices… | ||
| Copilot, Cursor, and Custom LLMs: Navigating the New .NET Developer Experience | Isaac Levin | 49m |
Compares real results: Copilot in Visual Studio gives deep MSBuild and debugger integration while Cursor Composer ships faster IDE-first features like… | ||
| From LLM Theory to Practical Agentic Implementations | Seth Juarez | 01h 00m |
Claim: LLMs are fundamentally a next-token predictor and have no independent agency; agency comes from the runtime. The demo uses the Prompty library and Draft… | ||
| Legacy Architecture Migration Patterns with DDD | Nick Tune | 01h 01m |
Shows concrete migration choices at PayFit: splits employee data across three legacy systems, migrates some features by moving writes first and others by… | ||
| Prompt-Jacking: The Rise of a New Supply Chain Risk | Kasimir Schulz, Kenneth Yeung | 01h 02m |
Findings show Cursor arbitrary code execution and data exfiltration by hiding commands in raw README comments and mermaid diagrams that trigger ls+curl to leak… | ||
| Use a Rust SAT Solver to play better pickleball! | Bob McNaughton | 54m |
Shows practical use of SAT/SMT tools: imports Rust SCIP crate (and mentions Z3) and models players as a two‑dimensional binary array plus a continuous… | ||
| Building AI That Learns and Adapts: A Case Study in MRI Diagnostics | Agata Chudzińska | 49m |
Project starts with 201 FCD MRIs plus ~100 normal and varied-pathology scans, preprocesses with brain tissue segmentation and two expert-inspired filters… | ||
| Everything you learned about SSL is deprecated | Todd H. Gardner | 51m |
Modern TLS replaces decades of complexity: TLS 1.3 cuts ~37 cipher suites to five and enforces perfect forward secrecy via Diffie‑Hellman so stolen long‑term… | ||
| "Looks Good to Me": A Practical Guide to Handling AI-Generated Code | Roman Zhukov | 01h 01m |
Open source projects see bots committing at scale (one bot hit nearly 100 repos); maintainers report up to 70% AI-assisted PRs in some flows and industry… | ||
| Doors of (AI)pportunity: The Front and Backdoors of LLMs | Kasimir Schulz, Kenneth Yeung | 53m |
Shows concrete exploits and defenses: prompt injection chains that stitch cultural references into SQL and image jailbreaks, an automated finder called… | ||
| Introducing the SQL MCP Server | Jerry Nixon | 56m |
Open source SQL MCP Server (aka data API builder) provides a JSON-configured, containerized API front end for SQL Server, Azure SQL, PostgreSQL, Cosmos DB and… | ||
| Threat Modeling Developer Behaviour: The Psychology of Bad Code | Tanya Janca | 58m |
Argues that many AppSec failures are predictable cognitive patterns (present bias, automation bias, bystander effect) and catalogs ten developer anti-patterns… | ||
| How to Break AI Systems (Before Someone Else Does) | Gary Lopez | 37m |
Large models fail because they treat instructions as data and are trivially jailbreakable; demos include hidden prompts in documents, agent goal manipulation… | ||
| Meditations on Code As Art | David Whitney | 46m |
Argues that code can be intentionally artistic by showing concrete artifacts: Tracey Emin’s My Bed sells for £2.54M as an example of human imprint, the Apollo… | ||
| Writing a Windows Kernel Driver in an hour or two | Pavel Yosifovich | 01h 02m |
Builds a small WDM driver and a userland client in about an hour that lets the client get powerful handles to protected processes (example: CSRSS) so it can… | ||
| On the benefits of non-intrusiveness | Patrice Roy | 01h 02m |
Compares intrusive OOP (public inheritance, interfaces) with non‑intrusive C++ techniques and shows concrete recipes: using traits and std::numeric_limits to… | ||
| Benchmarking and optimizing the Carbon compiler | Chandler Carruth | 01h 03m |
Project aims for an absurdly fast compiler and finds the real bottleneck in the lexer: identifier interning into a hash table. Uses Google Benchmark, Basil… | ||
| Expression-based Programming for Better Code? | Jason Turner | 01h 00m |
Claim: treating statements like expressions and moving each control block into its own function yields smaller, clearer C++ code that leverages RVO/NRVO, tail… | ||
| The Biggest Misconception of Computer Science | Alex Dathskovsky | 43m |
Argues that asymptotic Big‑O misleads because real performance is dominated by hardware: cache lines, latency, SIMD and branch predictors. Shows Dijkstra’s… | ||
| Lessons learned while coding my first functional type system in constexpr C++ | Daniel Nikpayuk | 55m |
A functional type system is implemented in constexpr C++ using continuation‑passing style controllers encoded as compile‑time byte arrays and a flat… | ||
| Postmodern Programming | Tony Van Eerd | 45m |
A loose manifesto for C++ code clarity that makes specific claims and shows working examples: a BlackBerry mouse‑pause context‑menu implementation that avoids… | ||
| Algorithms & Combinators | Conor Hoekstra | 57m |
Live demos show an AI-built toolkit (arraybox.dev, combinatory-logic pages and a vibecoded poll) that actually runs and maps algorithms across languages… | ||
| Boosting Software Efficiency : A Case Study of 100% Performance Improvement in ... | Gili Kamma | 01h 01m |
An embedded Linux unit with 32 MB RAM and C++/Qt codebase goes from frequent unexplained resets and data loss at ~5,000 meters to zero resets and support for… | ||
| Carbon: graduating from the experiment | Chandler Carruth | 01h 00m |
Carbon reports concrete progress: an implemented compiler and toolchain with deep C++ interop via Clang demonstrated live in Compiler Explorer, a working… | ||
| nothing matches | ||