Ankit Patel
How AI Models Get Smarter
#1about 2 minutes
How AI models are surpassing human experts
AI models are now exceeding human expert performance on comprehensive benchmarks like MMLU, which measures intelligence across various subjects.
#2about 5 minutes
The shift from labeled to unlabeled data training
The transformer architecture enabled a major shift from training on limited, human-labeled data to pre-training on vast amounts of unlabeled internet text using next-token prediction.
#3about 8 minutes
Refining models with post-training techniques
Pre-trained models are made useful for specific tasks like chatbots through post-training methods such as supervised fine-tuning and reinforcement learning from human feedback (RLHF).
#4about 3 minutes
Improving answer quality with reasoning models
Reasoning models improve accuracy by using test-time scaling, a process where the model prompts itself to double-check facts and logic before providing a final answer.
#5about 5 minutes
A practical workflow for AI application developers
Developers can build AI applications by starting with an API, using structured prompt engineering, and evaluating models in context rather than relying solely on benchmarks.
#6about 3 minutes
Implementing guardrails to secure your application
Protect your AI application from manipulation and misuse by implementing guardrails, detailed system prompts, and specialized guard models to enforce desired behaviors.
#7about 3 minutes
Building modular agentic applications with tools
Agentic applications use a modular architecture where each agent can use specific tools, often defined with natural language prompts, to perform complex tasks.
#8about 4 minutes
Q&A on model behavior and synthetic data
This Q&A covers why LLM responses are non-deterministic, how synthetic data is used for model distillation, and strategies for preventing hallucinations.
Related jobs
Jobs that call for the skills explored in this talk.
Wilken GmbH
Ulm, Germany
Senior
Kubernetes
AI Frameworks
+3
Matching moments
Featured Partners
Related Videos
Bringing the power of AI to your application.
Krzysztof Cieślak
AI: Superhero or Supervillain? How and Why with Scott Hanselman
Scott Hanselman
AI & Ethics
PJ Hagerty
WWC24 - Ankit Patel - Unlocking the Future Breakthrough Application Performance and Capabilities with NVIDIA
Ankit Patel
You are not an AI developer
Zan Markan
The shadows of reasoning – new design paradigms for a gen AI world
Jonas Andrulis
Staying Safe in the AI Future
Cassie Kozyrkov
Chatbots are going to destroy infrastructures and your cloud bills
Stanislas Girard
Related Articles
View all articles



From learning to earning
Jobs that call for the skills explored in this talk.

Forschungszentrum Jülich GmbH
Jülich, Germany
Intermediate
Senior
Linux
Docker
AI Frameworks
Machine Learning

Amazon.com Inc.
Senior
R
API
Unix
Perl
Ruby
+7

Amazon.com Inc.
Senior
R
API
Python
Matlab
Bootstrap
+4

10xEngineers
Remote
Senior
API
C++
Python
PyTorch
+4

FRG Technology Consulting
Intermediate
Azure
Python
Machine Learning



UL Solutions
Barcelona, Spain
Python
Machine Learning

Jack & Jill\u002FExternal ATS
Remote
Python
PyTorch
TensorFlow
Machine Learning
+1