Joinal Ahmed
AI Evangalist
Bengaluru, India
Actions
Joinal is an experienced Data Science professional with a interest on building solutions with quick prototypes, community engagements and influencing technology adoption. Brings strong technical understanding, experienced in managing cross functional teams of data scientists, data engineers, backend developers and SRE delivering end-to-end ML projects, recruiting and mentoring engineers across levels, streamlining ML & Data workflows for high performing teams, setting up best practices, and developing highly performant and reliable MLOps & Data platforms supporting end to end ml projects and analytics usecases.
Links
Area of Expertise
Topics
Effortless Scalability: Orchestrating Large Language Model Inference with Kubernetes
In the dynamic landscape of AI/ML, deploying and orchestrating large open-source inference models on Kubernetes has become paramount. This talk delves into the intricacies of automating the deployment of heavyweight models like Falcon and Llama 2, leveraging Kubernetes Custom Resource Definitions (CRDs) to manage large model files seamlessly through container images. The deployment is streamlined with an HTTP server facilitating inference calls using the model library.
This session will explore eliminating manual tuning of deployment parameters to fit GPU hardware by providing preset configurations. Learn how to auto-provision GPU nodes based on specific model requirements, ensuring optimal utilization of resources. We'll discuss empowering users to deploy their containerized models effortlessly by allowing them to provide a pod template in the workspace custom resource inference field. The controller dynamically, in turn, creates deployment workloads utilizing all GPU nodes.
On Device LLM with Keras and TF Lite
LLMs are very large in terms of storage, and generally consume a lot of computing power to run, which means they are usually deployed on the cloud and are quite challenging for On-Device Machine Learning (ODML) due to limited computational power on mobile devices.
In this session, attendees learn the techniques and tooling to build an LLM-powered app (using GPT-2 as an example model) with:
KerasNLP to load a pre-trained LLM
KerasNLP to finetune an LLM
TensorFlow Lite to convert, optimize and deploy the LLM on Android
Lets look at Terraform for MLOps
Terraform is one of the tools used by a lot of tech companies for managing their infrastructure. As an MLOps person do you really need to know about IaC? I will walk through the basics of Terraform and provision some GCP objects I used in my project.
In this session, let's see how we can use terraform to build GCP objects in an ML project.
Machine Learning Technical Debt : Using Kubeflow to Pay It Off Quickly
Kubeflow is a powerful and flexible MLOps platform . By using Kubeflow as the foundation of your MLOps platform, you can ensure the lifecycle of your ML projects are : uniformly managed, experiments are reproducible, prototyping is quick and you can ship your models into production without data scientists needing deep expertise y in infrastructure. We also saw how Kubeflow Pipelines enables developers to build custom ML workflows by easily “stitching” and connecting various components like building blocks using containerized implementations of ML tasks providing portability, repeatability and encapsulation.. Kubeflow makes it easy to deploy your machine learning models whether you’re running your code as notebooks or in Docker containers, Kubeflow allows you to focus on the model not the infrastructure.. Kubeflow also lowers the barrier to entry by providing a visual representation of the pipeline making it easy to understand and adopt for new users and provides an interactive UI to look at metrics and compare runs.
When Kubeflow is paired with Atrikko’s tools like Kale and Rok, data scientists can further focus on building their models and run experimentations while the heavy lifting of creating the docker containers, building the kubeflow pipeline, running hyperparameter tuning and deploying. With the models deployed using Kale and Rok, it ensures the reproducibility of experiments by storing a snapshot of the data, code and all other artifacts related to a run, which can be later used to replicate an experiment.
In this session we'll go over what an MLOps platform promises, how to deploy and Kubeflow in AWS ecosystem and ensure AI/ML governance.
The Hottest Programming language is English
Large Language models and Generative AI have taken the world by storm with their ability to generate coherent and diverse text, speech, images and more. These models, however, are not perfect and suffer from limitations such as data bias, lack of creativity, and low variability in output. In this talk, we will delve into the nuances of large language models and generative AI, exploring the impact of prompt engineering on the models' outputs. We will also discuss the strategies that can be employed to improve the models and tackle these limitations, including data augmentation, fine-tuning, and multi-model ensembling. By the end of this talk, attendees will gain a deeper understanding of the potential and limitations of these models, and the steps they can take to enhance their outputs and write prompts in a efficient way
Take Data to the Next Level With Graph Machine Learning
In this session, I will discuss why graph machine learning makes more sense than the traditional ML approach im certain ML usecases and show you how graph ML powers use cases like recommendation systems, fraud detection, and more. They will also teach you how to build a fraud detection solution powered by Neo4j and VertexAI in GCP, as well as how to deploy graph-based machine learning models on the cloud.
AI/ML Workloads Need Extra Security
The need for security is pervading all modern day systems. But given the growth in cloud machine-learning computing, which deals with extremely valuable data, and companies need to be paying particular attention to handling that data securely. The operation and maintenance of large scale production machine learning systems has uncovered new challenges which have required fundamentally different approaches to that of traditional software. The area of security in MLOps has seen a rise in attention as machine learning infrastructure expands to further critical use cases across industry. In this talk, we will discuss the key security challenges that arise in production machine learning systems, best practices and frameworks that can be adopted to help mitigate security risks in ML models, ML pipelines and ML services reinforcing SecOps into MLOps.
10 Things That Can Go Wrong w/ML Projects (and what you can do about it)
Machine learning practitioners are solving important problems every day. They're also experiencing a new set of challenges that are unique to ML projects.
This session will cover what to watch out for in terms of building a model; model accuracy; transparency and fairness; and MLOps.
The good news is that there are solutions. Attendees will hear about best practices and tools that will help address these issues.
ML Model in Production: Real-world example of End-to-End ML Pipeline with TensorFlow Extended (TFX)
Building the Machine Learning Model is just the first step. To monitor our predictions, offer alternatives that make our process scalable and adaptable to change, and maintain our model’s performance. In addition, it is important that we keep data regarding the execution of the pipelines, so that our processes are reproducible and that the error correction process is efficient. Using tools that support the process is essential to abstract the project’s complexity, making it more scalable and easier to maintain. The latest improvements of TensorFlow 2.0 are directed towards simplicity in model development and scaling. In this session we will look at how TFX Pipelines address DevOps and CI/CD requirements and compatibility with KubeFlow/Airflow/Apache Beam adds scalability into the mix.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top