Optimized AI Conference 2025
2025-04-14..2025-04-16
Atlanta, United States
[Optimized AI Conference 2025, Optimized AI Conference 2026]
| title | author | length |
|---|---|---|
| DIY PyTorch - How to Learn Deeply | Soumitra Pandit | 19m |
Explains how PyTorch’s implementation reduces deep learning to function composition and operator blocks, exposing a forward trace and an adjoint computational… | ||
| From Words to Data: Using LLMs to Turn Questions into SQL | Manohar Sai Jasti | 15m |
Claims that LLMs convert natural language to SQL in seconds, replacing hours or days, and demonstrates this end-to-end with a Jaffle Shop demo that actually… | ||
| Launching AI Pilots: Insights for Startups and Enterprise | Jason Quinley | 16m |
Many pilots fail because teams rush to deploy vague “AI solutions” without proving a real, desired outcome; the concrete advice is to map colleagues’… | ||
| Conversational Search with LLMs | Andrei Lopatenko | 24m |
Claim: LLMs let search move from single queries to long-term memory and multi-step, multi-channel flows. Concrete tools and patterns noted: a multi-tiered… | ||
| Commitment to Excellence: The Intersection of AI Business Innovation and Athletic Discipline | Sid Sriram | 12m |
Claims a rise from getting fired to building a seven-figure consulting business in under three years while competing as a national athlete, with named clients… | ||
| Unleashing the Future of AI Development | Logan Lawyer | 16m |
Claims that Dell Pro Max is a purpose-built AI dev workstation with Dell Pro Max systems shipping NVIDIA RTX Pro and Blackwell GPUs, preloaded with NVIDIA AI… | ||
| Beat GPT 4 and Build the Best AI For You: Mobilizing First-Party Data with Fine-Tuning | Donny Greenberg | 21m |
Claims that fine-tuning first‑party data lets teams beat off‑the‑shelf models with concrete wins: image classification shows 81% vs 95% when a tiny… | ||
| archive – Soumitra Pandit | Soumitra Pandit | 19m |
Explains how PyTorch implements adjoint computational graphs so each operator block carries its backward rule and a forward pass becomes a trace that can be… | ||
| archive – Sid Sriram | Sid Sriram | 12m |
A former data scientist gets fired, becomes an athlete, and in under three years builds a seven-figure consulting business while competing nationally. Daily… | ||
| archive – Manohar Sai Jasti | Manohar Sai Jasti | 15m |
Claims that large language models turn questions into near‑instant SQL answers and cut wait times from hours or days to seconds, demonstrated with a Jaffle… | ||
| archive – Jason Quinley | Jason Quinley | 16m |
Claim: many teams build “AI solutions” that don’t solve a real, desired problem. Concrete moves: validate the actual user need, ask for incentive plans of… | ||
| archive – Andrei Lopatenko | Andrei Lopatenko | 24m |
Twenty years of search experience frames the claim that conversational search needs long-term memory and multi-step, multi-channel flows. Empirical specifics… | ||
| Spark vs. Accelerated Compute: A Balanced Look at Data Processing Strategies | Giri Tatavarty | 19m |
Kroger runs 2,700 stores and $150B in annual revenue, so efficiency matters. Benchmarks show concrete wins: 1B-row sort: Spark 112s, Polars CPU 15s, Polars GPU… | ||
| Gen AI For Retail Data Management | Jan Luke Moyou | 16m |
Claims that generative AI automates data cleanup, attribute standardization, enrichment and governance and that multimodal LLMs translate and enrich product… | ||
| From Vibe-Coding to Fundraising: How AI Founders Stand Out in a Noisy Market | Devang Sachdev | 12m |
A concise six-question framework maps problem, market shift, impact, approach, team, and signals so clarity becomes a measurable output. Claims include raising… | ||
| Structuring AI Agents Projects for Small Business | Joshua Cazoe | 18m |
Claims come with specifics: about 300 client consultations, a push for specialization over generalization, and concrete tools like he gen and 11 labs, GPT‑3.5… | ||
| Revolutionizing Customer Care: Our Journey with LLM and RAG into Production | Clara Zang | 21m |
A two‑year build deploys an LLM + retrieval‑augmented generation stack into AT&T production with a secured vector store + semantic and hybrid search, chunking… | ||
| Solving Day 2 Problems: Hidden Hurdles to GenAI Success and How To Overcome Them | Shane Murray | 19m |
A field report that names specific failure modes and concrete fixes: a survey finds 100% of data professionals feel pressure to build GenAI and 91% have built… | ||
| Multi-LLM Routing Strategies for Gen AI | Ethan Ferdosi | 20m |
Twenty minutes condense experience with hundreds of startups into concrete patterns: use static routing for minimal latency where UI maps to one model, use… | ||
| No Code, No Problem: Building Powerful AI Agents with Just Your Words | Kimetha Dunn | 25m |
Claims that powerful no-code agents are buildable with plain prompts and off-the-shelf tools and then shows it: two demos — a custom GPT 'Strategy Sensei' with… | ||
| Unleashing the Future of AI Development | Logan Lawyer | 16m |
Claims a shift toward desk‑side heavy local compute using Dell Pro Max workstations with Blackwell GPUs and NVIDIA software, arguing local beats cloud for… | ||
| The Golden Triangle of Inference Optimization: Balancing Latency, Throughput, and Quality | Philip Kiely | 25m |
Baseten’s deck claims concrete wins: FP8 quantization on Lovelace/Hopper/Blackwell cuts bandwidth with preserved quality, distilling DeepSeek R1 671B into ~70B… | ||
| Applied AI & Data Analytics | Scott Radcliffe | 21m |
Describes a consulting-style practicum that solves a defined business problem in a 16-week cycle under client NDAs with weekly team and technical reviews… | ||
| Optimizing RAG with Semantic Caching & LLM Memory | Tyler Hutcherson | 14m |
Claims that semantic caching plus LLM memory cut cost and latency by reusing RAG outputs, with empirical numbers: ~30% question redundancy, ~60% cache hit… | ||
| Beyond Relevance: Using LLMs to Transform E-commerce Search into Product Discovery Engines | Yuanzheng Zhu | 20m |
Instacart uses large language models as a merchandising engine that generates complementary, substitute and bundle carousels layered under exact-match search… | ||
| Multi-Modal AI Optimization: Combining Text, Vision, and Speech Efficiently | Arpita Vats | 15m |
Argues that multimodal systems must fuse text, vision and speech with concrete architectures — early fusion, late fusion and a hybrid cross‑attention design —… | ||
| AI-Driven Precision Manufacturing | Balaji Chandrasekaran | 14m |
System-level inspection built around confocal microscopy plus precise XY staging rotates 17 mm balls and captures multifocal stacks, projects them to 2D, and… | ||
| Beat GPT 4 and Build the Best AI For You: Mobilizing First-Party Data with Fine-Tuning | Donny Greenberg | 21m |
Claims that fine-tuning on first‑party data reliably outperforms off‑the‑shelf LLMs, with concrete examples: Fashion‑MNIST: GPT‑4 ≈ 81% vs small ResNet/tree… | ||
| Double your Productivity at Work using these 3 GenAI Prompt Strategies | Patrick Hoeffel | 19m |
Claims a concrete productivity boost of "double" using three tactics and shows specifics that back it up: using long specific prompts, letting an LLM generate… | ||
| Knowledge Graphs and LLMs in Action | Allessandro Negro | 23m |
Claims that pairing knowledge graphs with LLMs creates an intelligent advisor that turns scattered structured and unstructured sources into a single… | ||
| archive – Shane Murray | Shane Murray | 19m |
Claims that moving gen AI past pilot reveals a persistent illusion of done and day‑two failures around reliability, observability, and trust. Cites survey… | ||
| archive – Joshua Cazoe | Joshua Cazoe | 18m |
Small businesses need concrete prescriptions: assess readiness, codify SOPs, clean and format data, and start with a minimum viable product rather than an… | ||
| archive – Ethan Ferdosi | Ethan Ferdosi | 20m |
Claims that multi-LLM setups outperform single models for evolving requirements and show three implementable patterns: static routing (UI components tied to… | ||
| archive – Clara Zang | Clara Zang | 21m |
A two‑year effort brings LLMs and a retrieval‑augmented generation stack into production at AT&T with a claimed company‑wide assistant that cuts agent search… | ||
| archive – Allessandro Negro | Allessandro Negro | 23m |
Claims that combining knowledge graph + LLM produces a verifiable, contextual advisor for time‑critical cases, using traffic camera feeds, car registries… | ||
| Revolutionizing Enterprise Workflows with Glean AI Agents | Bruce Schonk | 14m |
Claim: Glean’s platform unifies scattered knowledge without moving data, pulling from Salesforce Slack email G Drive Gong ServiceNow and the web to drive… | ||
| AI TechOps: Data and Infra for AI Foundation to Scale and Manage 1000s of Cohesive Agents | Kolin Konjura | 17m |
Claims a concrete blueprint for scaling to thousands of cohesive agents with orchestration layers that let third‑party agents interoperate (example: a host… | ||
| Improving RAG Systems - Key Challenges and Solutions | Naresh Dulam | 19m |
States that RAG remains fundamentally a search problem and lists concrete failure points: missing content, mis‑top ranking, incomplete extraction… | ||
| Memory-Augmented AI Agents: Enhancing Long-Term Reasoning and Context Awareness | Rahul Raja | 15m |
Claims that augmenting agents with external memory overcomes context windows (even references to a 4 million token context), reduces hallucination, and… | ||
| Building Transparent and Ethical Generative AI Models in High-Stakes Industries | Praneeth Reddy | 14m |
Claims specific mitigation steps for generative AI in healthcare and financial services include risk scoring by criticality and likelihood, concrete use‑case… | ||
| Unlocking the Full Potential of LLMs: Advanced Prompt Engineering Techniques | Lakshmi Narasannagari | 19m |
Claims that combining chain-of-draft with chain-of-thought compresses long reasoning (she shows 100-step runs summarized to a six-word output) while saving… | ||
| Why Fraud Detection, Sports Betting, and Bitcoin Prediction Look Alike | Brent Zucker | 23m |
Claims a unified ML lifecycle across fraud detection, sports betting, and Bitcoin prediction and gives concrete constraints: payments require 30 ms latency and… | ||
| Querying Databases with Function Calling | Connor Shorten | 23m |
Claims a function-calling API layer improves natural language database queries, builds a Weaviate query agent and transform agent, and benchmarks Gorilla LLM… | ||
| Leveraging AI for Faster Storage Access: A Graph-Neural-Network-Based Prefetcher | Faradawn Yang | 13m |
A graph neural network–based prefetcher converts raw addresses into deltas then builds a weighted graph with temporal edges plus low‑weight fully connected… | ||
| Overcoming the Small Nation Mindset - Competing Globally in AI | Learie Hercules | 32m |
Panel claims small nations can export AI by closing an imagination gap and reshaping identity, with concrete wins: a company using LLM-driven analytics with a… | ||
| Beyond RAG: Building Robust Knowledge Frameworks for Complex Multi-document Retrieval for Enterprise | Reshma Lal Jagadheesh | 21m |
Conventional RAG pipelines break on nested, multimodal, enterprise documents, so the talk presents concrete alternates: contextual retrieval (Anthropic paper… | ||
| Generative AI - A Pragmatic Business View | Bill Franks | 22m |
Practical claims and concrete patterns: generative models create new records so use them for synthesis, summarization, translation and extraction; prefer RAG… | ||
| Revolutionizing Client Inquiry Management with Gen-AI | Lin Zhang | 17m |
Claims a lightweight, production-ready system that uses few-shot prompting and an LLM endpoint to tag inquiries by sentiment and high/medium/low urgency… | ||
| Creating an AI Culture in the Enterprise | Mark Tabladillo | 21m |
Claim: a three‑pillar blueprint (cross‑functional communities, hands‑on education, alignment to production) scales. Evidence: an internal program grows from… | ||
| Leveraging GenAI in Customer Service to Improve Customer Experience and Reduce Operational Costs | Shankar Krishnan | 19m |
Claims include virtual agents and real-time agent assist cutting human-handled call cost from $5–6 to $2–3, handling a support volume that grows ~30–40% per… | ||
| AI For Executives: What You Need to Know | Learie Hercules | 29m |
Claims that AI collapses time and effort to insight, delivering LLM-driven analytics platform outcomes for supply chain customers; demonstrates a working ROI… | ||
| From Chaos to Control: Building Enterprise-Grade LLM Agents That Actually Work | Vivek Raju Muppalla | 25m |
Claims that 99% per-step accuracy still yields ~18% failure over 20 steps and 95% yields nearly two-thirds failures drive concrete fixes: use the largest model… | ||
| Building Startups 3.0 | Kyle Maloney, Devang Sachdev | 23m |
Argues concrete claims: AI unfolds in three waves—enterprise productivity, consumer agents, then larger deep tech opportunities—and that Startups 3.0 are lean… | ||
| Practical AI: Driving Tangible Value in the Enterprise | Bill Franks | 27m |
Panelists name concrete wins and failures: a computer vision system at Allegiance Stadium recycling demo uses existing cameras, bounding boxes and haptics… | ||
| The Interconnectedness of Artificial Intelligence and Organizational Behavior | Laura Edell | 23m |
Claims a measurable metric called knowledge priors — the rate people convert life experience into skill — and applies algorithmic information theory to… | ||
| Avoiding RAG Sprawl in the Enterprise | Amr Awadallah | 34m |
Argues that enterprise RAG often produces fragmented, duplicative pipelines that break governance, security, cost and scaling; contrasts DIY stacks using… | ||
| The World of Vertical AI Agents | Hamza Farooq | 26m |
Claims most agent demos are LLM wrappers with little production value and pushes for narrow, domain-focused systems. Shows Manis AI on Claude as a quick… | ||
| The Promise of AI in Education | Bill Franks | 25m |
Panelists state concrete shifts: courses add theoretical advanced ML and RAG and agentic frameworks; LLMs already solve many assignment problems and generate… | ||
| Raising Venture Capital in the Gen AI Era | Optimized AI Conf | 24m |
Funding shifts from hype to results with chatGPT made AI mainstream driving splashy rounds even as money tightens; early-stage investors write roughly average… | ||
| Humanity and AI | Beverly Wright, Eileen Oriel | 26m |
Panelists report concrete changes: ambient AI scribes now automate transcription and propose personalized treatment plans, people rely on ChatGPT Plus with… | ||
| How to Make Useful AI Agents | Sam Partee | 17m |
Claims a production-ready stack: Arcade.Dev infrastructure for OAuth, multi-tenant tool runtime, and cached user tokens. Demonstrations include ambient Slack… | ||
| Making Deep Learning Easier - Updates on Keras | Monica Song | 19m |
Keras shows concrete, usable changes: model.fit now trains the same model code across JAX TensorFlow and PyTorch backends, with multi‑device sharding defined… | ||
| Large-scale Enterprise Co-pilots - Making AI Useful for Business | Dona Sarkar | 27m |
Claims that matter: Microsoft already ships 110 co-pilot products and defines an agent as an LLM that takes inputs and outputs actions, not just chat… | ||
| The Bucket Brigade: Why S3 is the Missing Piece in your AI Architecture | Jacopo Tagliabue | 23m |
Combining Iceberg with simple Python functions run in the cloud turns S3 from a dumb blob store into tables on S3 that support pushdowns and time travel, plus… | ||
| AI Engineering Anti-patterns | Chip Huyen | 32m |
Talk lists concrete anti‑patterns and shows what actually fails: claims of “30% energy savings” collapse on user behavior and simpler convex optimization would… | ||
| Building at the Speed of Thought | Paige Bailey | 21m |
Claims a near‑term model that "writes code better than any human" and shows concrete demos: Gemini 2.5 Pro analyzing an 88,000‑token five‑minute video… | ||
| nothing matches | ||