Abstracts
Applications to speak. These go through review and a decision.
| Tags | Speaker | Ratings | Files | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Docs That Answer Back: Retrieval-Grounded Documentation Sites | Pending | Approved | — | Abstract | — | A 10-minute tour of turning a static docs site into one that answers questions with citations, stays honest when it doesn't know, and costs under $50/month to run. Live demo, real failure cases, and a checklist you can apply to your own docs this week. | — | — | — | — | — | Sswyx+kms-0033-speaker@ai.engineer | Sswyx+kms-0033-speaker@ai.engineer | — | — | — | — | — | |
| Lightning: Agents in Production Q&A | Accepted | Approved | Yes | Abstract | — | A rapid-fire Q&A lightning talk on running AI agents safely and reliably in production, covering observability, guardrails, and rollback strategies. | — | — | — | — | — | Ssbek-speaker2@example.com | Ssbek-speaker2@example.com | — | — | — | — | — | |
| Your AI Pair Programmer Is Lying to You: Verification Patterns That Scale | Pending | Approved | — | Abstract | — | Code generation is easy; trusting it is hard. This session covers verification patterns for AI-generated code — property tests, mutation coverage, snapshot judges, and CI gates — with data from 18 months of running them on a 200-engineer codebase. Includes what we stopped doing because it didn't catch anything. | — | — | — | — | — | Ssbek-speaker2@example.com | Ssbek-speaker2@example.com | — | — | — | — | — | |
| Docs That Answer Back: Retrieval-Grounded Documentation Sites | Accept Queue | Approved | — | Abstract | — | A 10-minute tour of turning a static docs site into one that answers questions with citations, stays honest when it doesn't know, and costs under $50/month to run. Live demo, real failure cases, and a checklist you can apply to your own docs this week. | — | — | — | — | — | AOAda Okonkwo | AOAda Okonkwo | — | — | — | — | — | |
| Taming 40-Minute CI: Incremental Builds at Monorepo Scale | Accepted | Approved | Yes | Abstract | — | Our monorepo CI took 40 minutes on a good day. This talk walks through how we cut it to 6 minutes with content-addressed caching, remote execution, and a test-selection model — including the two migrations that failed first. You'll leave with a decision framework for which incremental-build investments pay off at which repo sizes, and the graphs to convince your platform team. Updated: now includes 2026 benchmark data. | — | — | — | — | — | AOAda Okonkwo | AOAda Okonkwo | — | — | 3.63 (2) | — | — | |
| Your AI Pair Programmer Is Lying to You: Verification Patterns That Scale | Accept Queue | Approved | — | Abstract | — | Code generation is easy; trusting it is hard. This session covers verification patterns for AI-generated code — property tests, mutation coverage, snapshot judges, and CI gates — with data from 18 months of running them on a 200-engineer codebase. Includes what we stopped doing because it didn't catch anything. | — | — | — | — | — | AOAda Okonkwo | AOAda Okonkwo | — | — | 5.00 (1) | — | — | |
| Full Journey Demo Talk (Round 2) | Accepted | Approved | Yes | Abstract | — | A clean re-run of the full submit-to-publish journey, with the Organiser assigning the abstract to a committee, a Reviewer blind-scoring it, and the Organiser deciding, notifying, scheduling and publishing based on that score. | — | — | — | — | — | Ffull-journey-demo-2@example.com | Ffull-journey-demo-2@example.com | Oct 11, 2026, 12:00 AM | Room B | 4.50 (1 of 4) | — | — | |
| Reviewer-First Journey Talk | Accepted | Approved | Yes | Abstract | — | A demo submission created to walk the proper submit-review-accept-schedule-publish journey, with the reviewer step included this time. | — | — | — | — | — | Rreviewer-first-demo@example.com | Rreviewer-first-demo@example.com | Oct 10, 2026, 11:00 PM | Room B | 4.00 (1 of 4) | — | — | |
| Full Journey Walkthrough Talk | Accepted | Approved | Yes | Abstract | — | A live demo walkthrough talk created to show the full submit-to-publish journey end to end. | — | — | — | — | — | Wwalkthrough-demo@example.com | Wwalkthrough-demo@example.com | Oct 10, 2026, 10:00 PM | Room A | — | — | — | |
| Claude - Event | Pending | In Review | — | Abstract | — | Test | — | — | — | — | — | Ffarishussain021@gmail.com | Ffarishussain021@gmail.com | — | — | — | — | — | |
| E2E Journey Test Talk | Accepted | Approved | Yes | Abstract | — | A test submission created to walk the full submit-to-publish journey end to end. | — | — | — | — | — | Ee2e-tester@example.com | Ee2e-tester@example.com | Oct 10, 2026, 8:00 PM | Room A | — | — | — | |
| Closing Remarks | Pending | Approved | — | Abstract | AI-039 | Closing thoughts, thanks, and what we would like to see submitted next year. | Talk (30 min) | Advanced | English | Evaluation | Open SourceSecurity | SRSofia Rossi | SOSam Organiser | — | — | — | — | 0.9 | |
| Opening Keynote: Why Now | Pending | Approved | — | Abstract | AI-040 | Why this conference, why now, and what we hope you take away from the next two days. | Talk (30 min) | Beginner | English | Product | Open Source | WCWei Chen | SOSam Organiser | — | — | — | — | 1.2 | |
| Panel: The Ethics of Autonomy | Pending | Approved | — | Abstract | AI-038 | Autonomy raises questions the industry has been deferring. A panel on accountability, disclosure and the decisions we are quietly making on users' behalf. | Talk (30 min) | Intermediate | English | Infrastructure | CostSecurity | IKIdris Khan | SOSam Organiser | — | — | — | — | 0.2 | |
| Panel: Buy, Build or Wait | Draft | Approved | — | Abstract | AI-037 | Three practitioners argue about when to buy, when to build and when to wait. Audience questions throughout; the panel has been asked not to agree. | Talk (30 min) | Beginner | English | Applied AI | AgentsCost | NFNaomi FischerWCWei Chen | SOSam Organiser | — | — | — | — | 0.1 | |
| Workshop: Production RAG in Half a Day | Withdrawn | Approved | — | Abstract | AI-035 | Half a day taking a RAG prototype to something you would put in front of customers: chunking, retrieval quality, evaluation and the deployment checklist. | Talk (30 min) | Intermediate | English | Evaluation | Open SourceRAG | PRPriya Raman | SOSam Organiser | — | — | — | — | 2.1 | |
| Lightning: Five Prompts That Changed Our Product | Draft | Approved | — | Abstract | AI-036 | Five prompts that measurably changed a product metric, what they replaced, and why each one worked. Five minutes, no slides after the first. | Workshop (120 min) | Advanced | English | Product | AgentsRAG | TBTomas Berg | SOSam Organiser | — | — | — | — | 1.4 | |
| Workshop: Building Your First Evaluator | Withdrawn | Approved | — | Abstract | AI-034 | A hands-on session building an evaluator from scratch. Bring a laptop; you will leave with a working harness scoring your own outputs against a labelled set you build in the room. | Talk (30 min) | Beginner | English | Infrastructure | Open SourceSecurity | LMLuis Moreau | SOSam Organiser | — | — | — | — | 1.3 | |
| Debugging Agent Loops | Declined | Approved | Yes | Abstract | AI-030 | Agent loops fail in ways stack traces do not capture. The trace format we adopted, how we replay a failing run locally, and the visualisation that made the loops legible to people who did not write them. | Talk (30 min) | Advanced | English | Infrastructure | Open SourceRAG | SWSam Whitfield | SOSam Organiser | — | — | — | — | 0.4 | |
| Batching, Queueing and Backpressure | Declined | Approved | Yes | Abstract | AI-028 | Queue depth, batch size and tail latency pull against each other. What we learned running a shared inference tier, and the backpressure design that stopped one team's spike from taking down everyone. | Talk (30 min) | Beginner | English | Product | Cost | KMKofi Mensah | SOSam Organiser | — | — | — | — | 0.5 | |
| Migrating Off a Single Vendor | Declined | Approved | Yes | Abstract | AI-029 | Moving off a single provider without a rewrite. The compatibility layer, the differences that leaked through anyway, and an honest cost comparison after the dust settled. | Workshop (120 min) | Intermediate | English | Applied AI | Security | EPElena PetrovaADAmara Diallo | SOSam Organiser | — | — | — | — | 1 | |
| Semantic Caching in Practice | Decline Queue | Approved | — | Abstract | AI-026 | Caching on meaning rather than exact match, and the ways that goes wrong. Threshold tuning, the confidently-wrong cache hit, and how we detect when the cache is degrading answer quality. | Talk (30 min) | Intermediate | English | Infrastructure | RAG | JHJonas Halvorsen | SOSam Organiser | — | — | 4.50 (1 of 4) | — | 1.5 | |
| The Hidden Cost of Context Windows | Decline Queue | Approved | — | Abstract | AI-027 | Long context is not free and not always better. Measured cost and quality across context lengths on a real workload, plus the retrieval strategy that beat simply pasting everything in. | Talk (30 min) | Advanced | English | Evaluation | Agents | MCMira Castellanos | SOSam Organiser | — | — | — | — | 2.4 | |
| Open Weights in the Enterprise | Decline Queue | Approved | — | Abstract | AI-025 | Open-weight models inside a company with a compliance function. Licensing, hosting, evaluation against hosted alternatives, and the security questions that decided it for us. | Talk (30 min) | Beginner | English | Applied AI | Open Source | GAGrace AdeyemiKMKofi Mensah+1 | SOSam Organiser | — | — | 2.75 (1 of 4) | — | 2.7 | |
| Designing for Model Deprecation | Accepted | Approved | Yes | Abstract | AI-024 | Every model you build on will be deprecated. Abstraction boundaries that survive a swap, the regression suite that makes a migration a day instead of a quarter, and what not to abstract. | Talk (30 min) | Advanced | English | Product | Open SourceSecurity | WCWei Chen | SOSam Organiser | — | — | 3.00 (1 of 4) | — | 0.7 | |
| Scaling Human Review | Accepted | Approved | Yes | Abstract | AI-022 | Human review is the bottleneck nobody budgets for. Queue design, reviewer fatigue, inter-rater agreement, and the tooling changes that tripled throughput without changing headcount. | Workshop (120 min) | Beginner | English | Infrastructure | AgentsCost | IKIdris Khan | SOSam Organiser | Oct 11, 2026, 6:00 PM | Room B | 4.25 (1 of 4) | — | 2.9 | |
| What We Learned Serving 10B Tokens | Accepted | Approved | Yes | Abstract | AI-023 | Ten billion tokens through one API. Capacity planning, the failure modes that only appear under sustained load, and the three architectural decisions we would make differently. | Talk (30 min) | Intermediate | English | Evaluation | CostSecurity | SRSofia Rossi | SOSam Organiser | Oct 11, 2026, 6:30 PM | Room B | 2.50 (1 of 4) | — | 2.6 | |
| Small Models, Big Wins | Accepted | Approved | Yes | Abstract | AI-020 | A 3B model beats a frontier model on our task, at a fraction of the cost. How we found that out, how we validated it, and the distillation pipeline that keeps it current as the task drifts. | Talk (30 min) | Intermediate | English | Product | Open Source | TBTomas Berg | SOSam Organiser | Oct 10, 2026, 5:00 PM | Room A | 2.75 (1 of 4) | — | 1.6 | |
| Retrieval Quality Is a Data Problem | Accepted | Approved | Yes | Abstract | AI-021 | Retrieval failures are almost never the embedding model. Chunking, metadata, document freshness and access control — a walk through the four data problems that caused most of our bad answers. | Talk (30 min) | Advanced | English | Applied AI | AgentsRAG | NFNaomi FischerWCWei Chen | SOSam Organiser | Oct 12, 2026, 5:00 PM | — | 3.25 (1 of 4) | — | 1 | |
| Testing Nondeterministic Pipelines | Accepted | Approved | Yes | Abstract | AI-018 | You cannot assert equality on a generative pipeline. Property-based checks, golden sets, statistical gates in CI, and how we stopped flaky tests from training the team to ignore red builds. | Talk (30 min) | Advanced | English | Infrastructure | CostSecurity | LMLuis Moreau | SOSam Organiser | Oct 10, 2026, 5:00 PM | Room A | 2.81 (4) | — | 1.4 | |
| The Case Against Chatbots | Accepted | Approved | Yes | Abstract | AI-019 | Chat is a lazy default. Three products where we replaced a conversational interface with something more constrained, the adoption change that followed, and the one case where chat genuinely won. | Talk (30 min) | Beginner | English | Evaluation | Security | PRPriya Raman | SOSam Organiser | Oct 10, 2026, 5:00 PM | Room A | 2.88 (4) | — | 0.6 | |
| From Notebook to Production in a Week | Accepted | Approved | Yes | Abstract | AI-017 | A team of two took an internal prototype to production in six days. What we skipped deliberately, what that cost us later, and the checklist we now use to decide which corners are safe to cut. | Talk (30 min) | Intermediate | English | Applied AI | AgentsCost | AOAda OkonkwoTBTomas Berg | SOSam Organiser | Oct 11, 2026, 4:30 PM | Main Hall | 3.13 (4) | — | 0.6 | |
| Latency Budgets End to End | Accepted | Approved | Yes | Abstract | AI-016 | Latency is a chain, not a number. Tracing a request from browser to model and back, finding the three places we were wasting a second each, and the budget we now hold teams to. | Talk (30 min) | Beginner | English | Product | AgentsRAG | ADAmara Diallo | SOSam Organiser | Oct 12, 2026, 9:00 PM | Main Hall | 2.25 (4) | — | 2.2 | |
| Multi-tenant Isolation for AI Workloads | Accept Queue | Approved | — | Abstract | AI-014 | Serving many customers from shared model infrastructure without leaking context between them. Isolation boundaries, noisy-neighbour throttling, and the audit story that survived a security review. | Talk (30 min) | Intermediate | English | Infrastructure | Open SourceSecurity | SWSam Whitfield | SOSam Organiser | Oct 10, 2026, 4:00 PM | Main Hall | 3.25 (4) | — | 1.7 | |
| Guardrails That Do Not Annoy Users | Accepted | Approved | Yes | Abstract | AI-015 | Safety filters that block real users are a product bug. We describe the layered approach we use, how we measured false positives against actual traffic, and the categories where we deliberately accept more risk. | Workshop (120 min) | Advanced | English | Evaluation | Open SourceRAG | YTYuki Tanaka | SOSam Organiser | Oct 10, 2026, 4:00 PM | Main Hall | 3.13 (4) | — | 1.4 | |
| Caching Strategies for Token Spend | Accept Queue | Approved | — | Abstract | AI-013 | Semantic and exact caching in front of a production assistant. Hit rates by traffic type, the staleness bugs that followed, and the invalidation strategy we ended up with after two rewrites. | Talk (30 min) | Beginner | English | Applied AI | CostSecurity | EPElena PetrovaADAmara Diallo | SOSam Organiser | — | — | 2.31 (4) · 1.25–3.25 | — | 0.5 | |
| When to Say No to an Agent | Accept Queue | Approved | — | Abstract | AI-012 | Not every problem deserves an agent. A decision framework drawn from six internal projects, three of which we cancelled, and what the cancelled ones had in common that we did not see at the time. | Talk (30 min) | Advanced | English | Product | Agents | KMKofi Mensah | SOSam Organiser | — | — | 3.13 (4) | — | 2.9 | |
| Building an Internal Model Gateway | Accept Queue | Approved | — | Abstract | AI-011 | We put every model call behind one internal service. This covers routing, per-team quotas, the audit trail that made compliance stop worrying, and the migration that let us swap a vendor in an afternoon. | Talk (30 min) | Intermediate | English | Evaluation | AgentsRAG | MCMira Castellanos | SOSam Organiser | May 14, 2027, 5:00 PM | Overflow Room | 2.88 (4) | — | 3 | |
| Structured Output Without Tears | Accept Queue | Approved | — | Abstract | AI-010 | Getting reliable JSON out of a model that would rather write prose. Grammar-constrained decoding, schema repair, and why we stopped asking the model to apologise for malformed output and started making it impossible. | Talk (30 min) | Beginner | English | Infrastructure | Open Source | JHJonas HalvorsenAOAda Okonkwo | SOSam Organiser | May 14, 2027, 4:00 PM | Overflow Room | 2.44 (4) · 1.75–3.75 | — | 3 | |
| Edge Inference on a Budget | Pending | Approved | — | Abstract | AI-009 | Running models close to users without a GPU budget. Quantisation choices, the memory ceiling on commodity edge hardware, and an honest account of the quality we traded away and where it turned out to matter. | Talk (30 min) | Advanced | English | Applied AI | Open SourceSecurity | GAGrace AdeyemiKMKofi Mensah | SOSam Organiser | — | — | 2.13 (4) | — | 1.2 | |
| Streaming UIs for Slow Models | Pending | Approved | — | Abstract | AI-008 | When the model takes eleven seconds, the interface is the product. Streaming, skeletons, optimistic rendering and the moment users decide something is broken — with the abandonment numbers that changed our minds. | Workshop (120 min) | Intermediate | English | Product | Cost | WCWei Chen | SOSam Organiser | — | — | 2.88 (4) · 1.75–4.25 | — | 2.8 | |
| Prompt Injection in the Wild | Pending | Approved | — | Abstract | AI-007 | A field report on prompt injection attempts against a public-facing assistant: what was tried, what worked, and which mitigations were theatre. Includes the payload that got past three layers of filtering. | Talk (30 min) | Beginner | English | Evaluation | AgentsCost | SRSofia Rossi | SOSam Organiser | — | — | 3.25 (3) | — | 1.5 | |
| Observability for Nondeterministic Systems | Pending | Approved | — | Abstract | AI-006 | Traditional observability assumes the same input gives the same output. We cover the tracing schema we settled on, how we sample when every request is unique, and how to alert on quality drift without drowning in false positives. | Talk (30 min) | Advanced | English | Infrastructure | RAG | IKIdris Khan | SOSam Organiser | — | — | 3.31 (4) · 2.25–4.5 | — | 2.6 | |
| Fine-tuning Is Not the Answer | Pending | Approved | — | Abstract | AI-005 | Fine-tuning is the first thing people reach for and usually the wrong one. We compare it against retrieval, prompt work and routing on the same three tasks, with the training costs and the maintenance burden included honestly. | Talk (30 min) | Intermediate | English | Applied AI | Open SourceRAG | NFNaomi FischerWCWei Chen | SOSam Organiser | — | — | 4.00 (4) | — | 0.6 | |
| The Cost Curve of Inference | Pending | Approved | — | Abstract | AI-004 | Inference costs do not scale the way finance expects. A breakdown of where our spend actually went across a year — prefill versus decode, cache hit economics, and the surprising fraction consumed by retries and abandoned streams. | Talk (30 min) | Beginner | English | Product | Open SourceSecurity | TBTomas Berg | SOSam Organiser | — | — | 2.63 (4) | — | 1.9 | |
| Vector Databases at a Billion Rows | Pending | Approved | — | Abstract | AI-002 | Everything is fast at ten million rows. We walk through what actually broke between one hundred million and a billion: index build times, memory-mapped segment churn, and the recall cliff nobody warns you about when you quantise too aggressively. | Talk (30 min) | Intermediate | English | Infrastructure | Agents | LMLuis Moreau | SOSam Organiser | — | — | 3.06 (4) | — | 1.2 | |
| Evaluating RAG: Beyond Vibes | Pending | Approved | — | Abstract | AI-003 | Most RAG evaluation is a demo and a feeling. We describe the offline harness we built, why we abandoned answer-similarity scoring, and how a small hand-labelled set of two hundred questions caught regressions our automated metrics happily approved. | Talk (30 min) | Advanced | English | Evaluation | CostSecurity | PRPriya Raman | SOSam Organiser | — | — | 3.00 (4) · 1–5 | — | 2.8 | |
| Shipping LLM Agents Without Losing Sleep | Pending | Approved | — | Abstract | AI-001 | We ran agents in production for eighteen months and most of what we believed at the start was wrong. This covers the retry semantics, the budget guards and the three incidents that shaped our current design, including the one that took a weekend to unpick. | Workshop (120 min) | Beginner | English | Applied AI | RAG | AOAda OkonkwoTBTomas Berg | SOSam Organiser | — | — | 3.31 (4) | — | 1.1 |
1–48 of 48
48 abstracts matching the current filters↑↓move⇧↑select rangespacetoggle⏎openescclear
1 / 1