Laisha Wadhwa
Head of Devrel @Pesto Tech | Ex @Goldman Sachs | Podcast Host| WTM Ambassador| Tech Speaker| Mentor
Delhi, India
Actions
I'm Laisha Wadhwa! As the head of Devrel at Pesto tech, I get to engage with brilliant minds in the tech community, fostering connections and helping developers reach new heights.
Before joining Pesto, I was working as a Senior Software Developer at Goldman, where I worked on building consumer banking products and dealt with security and risk aspects of fintech. My passion for Data and Machine Learning led me to explore innovative ways to leverage these technologies for solving real-world problems.
However, my journey doesn't stop there. I've been completely captivated by the possibilities that the world of Web3 brings to the table.
One of my proudest achievements is the development of RadarFi, a cutting-edge security product that helps protect users in the digital realm. My drive to create tech for social good has been a guiding force throughout my career, and I'm committed to using AI and other groundbreaking technologies to usher in the next generation of revolution.
Links
Area of Expertise
Topics
Don't just learn tech, learn to build things
Learning to code has been gaining a lot of ground lately. But coding is not the new literacy. It's not always about learning the technology, its about using those skills to build things, products, Prototypes etc.
As a student during college I was big on learning a plethora of technologies, reading about it, watching all kinds of tutorials, but I had not ventured into practically using those skills beyond my academic curriculum until my second year of college.
I stumbled upon the ideas of hackathons in 2018 and that's when my actual journey in tech began.
Participating in my first national hackathon was an eye opener for me when I understood that learning to code is not the be all and end all of the tech industry.
Learning to build things shaped me as a developer, as a product designer and as a team player too.
In this talk I will share my journey of adopting the ideas of learning to build. I will be shedding light on the path that I chose, to be where I am today, of how the extra efforts that I put at the right time at the right place helped me accelerate my career, build a name for myself and also build products. Building things makes you aware of your own capabilities and also bolsters your confidence. It teaches you smart management, team dynamics ad much more. All these skills not only helps one drive their career better but also contributes to becoming a comprehensive developer
Being a woman in tech, the journey till here wasn't easy. Just like me there are so many women who venture into this field with the idea of learning to code. I want to bring out the message: "Don't just learn technology, learn to build things" to the wider audience and motivate them to go to that hackathon, participate in that bootcamp, go talk about their projects and takeaway some inspiration and positive affirmation from my journey which in turn powers many more women and people in general.
I was inspired and motivated by so many wonderful folks in tech to tread on this journey and the lessons that I have learnt on the way have helped me build a path for myself. Through the talk I want to give back to the community that helped me reach here and bring out the idea of learning to build as a skill that each one of us should have.
Why Machine Learning Models Crash And Burn In Production
A model’s accuracy will be at its best until you start using it. It then deteriorates as the world it was trained to predict in changes. MLOps pipelines need to prepare for change!
Join in to learn the best practices for ML pipeline automation to achieve continuous delivery of model prediction service using Streamlit, Google K8s & Git
Why Do ML Models Fail?
Implementing ML in a production environment ≠ deploying your model as an API for prediction. It means deploying an ML pipeline that can automate the retraining and deployment of new models.
Learn how you can use CD & automation pipelines in your ML pipelines to have:
- Active performance monitoring
- Understanding the cadence to retrain your production models
- Learn about templates for setting up robust ML pipelines with right triggers.
During the talk, I’ll cover Level 0 & Level 1 of automation for ML pipelines & discuss implementations of CI, CD and CT using Google's K8s and vortex AI
Unlocking Efficiency: End-to-End ML with Vertex AI
Discover how to train custom models using AutoML, evaluate their accuracy, & deploy them with Vertex AI + integration with GCP & utilization of KFP SDK.
This session dives deep into constructing an end-to-end AutoML workflow using Vertex Pipelines. We'll explore the use of AutoML Tabular to train models with structured data, using the Dry beans dataset as an eg, a classification task predicting bean types based on characteristics.
While prebuilt Google Cloud Pipeline Components play a significant role in the pipeline, I'll also introduce a Python function-based custom component for model evaluation & metrics visualization. Will also showcase the versatility of Vertex Pipelines in simplifying complex ML workflows.
What Will You Learn:
- Learn how to construct end-to-end ML workflows with a focus on AutoML training.
- Explore the integration of AutoML Tabular with real-world datasets.
- Unlock the potential of custom components for advanced model evaluation and visualization.
Autonomous Vehicles See More With Thermal Imaging: Multi-modal thin cross section Object Detection
In the era of AI, there's a renewed focus on making autonomous cars a reality. However, there are many fallbacks like spurious identification of pedestrians as a piece of paper, ramming into bicycles due to missed identification. But the problem can be solved using thermal Images instead of RGB. This talk revolves around how I leveraged thermal images to enhance the vision of autonomous cars!
I'll be talking about Firefly - a thermal image-based object detection module for autonomous cars that I built during the Mercedes Benz Digital Hackathon this year. My learnings and thoughts on why Why Autonomous Vehicles Need Thermal Cameras and why the vehicle industry should no longer remain cool to using thermal imaging.
Outline
Introduction [5 Minutes]
Who am I?
Weren't autonomous cars doing good already?
What's the buzz around tech4autonomous.
The current problem of object detection.
Real-World examples of fatal consequences. (Uber and Tesla).
Thermal dataset- FLIR dataset - Theory [10 minutes]
Why do Thermal Cameras fill in the Autonomous Vehicle Sensor Gap?
Why use it?
Understanding why it's better than existing LIDAR & camera-based solutions.
Processing thermal images.
What to do and what not to!
Generating Thermal images (image translation - CycleGAN)
Data Preparation [6 minutes]
How do we use the images?
Using CYCLE GAN to generate more data for model enhancement.
Building an object detection model in Tensorflow [9 minutes]
Pretrained RGB network on PVOC - why it failed.
Experimenting with YOLOv3 and RCNN - practical results
Why YOLOv3? Learnings from failures.
Seeing RGB - thermal in Action [1-minute demo] [4 minutes]
Comparing RGB outputs vs Thermal Outputs. (Let's see some numbers)
Challenges faced.
Future scope for FireFly (Open-sourced for contribution: [https://github.com/laishawadhwa/firefly])
Key Takeaways, Q&A
What's the buzz around autonomous cars?
In recent times, our reliance on automation has been increasing exponentially. This is specifically evident in the automotive industry, where we see an aggressively large number of software components being used in automotive cars and vehicles.
This is done with the hope of providing better safety, comfort, and assistance to the driver, while also improving the experience of passengers. In the last few years, there has been extensive growth in Artificial Intelligence (AI) research.
The automotive industry has been increasing its use of AI technologies, as these have the potential to enhance driving experiences.
The problem
Recognition of thin cross-section objects changing direction (e.g. a cyclist, pedestrian) is still relatively difficult. One would argue- we have state of the art Object detection models and LIDAR sensors!! But they don't solve the real problem.
Bicycles are generally considered “the most difficult detection” problem that autonomous vehicle systems face due to their thein cross-section.
Unmanned vehicles cannot rely on visible images while navigating in cloudy weather/low sunlight or during the night.
The solution - Content Theory
We Use thermal Images!
The FLIR dataset: It helps detect and classify pedestrians, bicyclists, animals, and vehicles in challenging conditions like total darkness, fog, smoke, inclement weather, and glare.
Detection range: 4x farther than typical headlights.
Preprocessing Data: Enhance edges using a Butterworth high pass filter. Handling class imbalance across training and validation sets using the oversampling technique. Narrowing down to only three categories(People, Bicycles, and Cars).
FLIR dataset isn't sufficient to train an accurate model! - USe CYCLE GAN to generate more data.
Leveraging transfer learning by using pre-trained convolution weights from ResNet pre-trained on FLIR ADAS and Pascal VOC.
Visiting the network architecture to understand why YOLOv3 works!
Comparing the shortcomings of RGB-based object detection with thermal images based detection on real-world videos.
Discussion on results and challenges that need to be addressed.
The future scope of handling the problem of occlusion can be handled by predicting the optical flow/motion trajectories of each agent in the video.
What's in store for you?
The talk will be very interactive and has a lot of fun scenarios to explore. All the math enthusiasts will sure have a lot of fun understanding the proposed network architectures. The scenarios I'll be dealing with will be very relatable and something anyone who drives would have faced. There'll be plenty of practical examples to understand and it will be quite an interesting use case to explore for the audience.
Key Takeaways:
The talk will give you a new perspective on working with thermal images and autonomous driving applications. You'll have enough information and methods the to program object detection modules for autonomous cars and will also be able to research the topic further with little or no help.
You'll get to explore new use- cases using GAN's, YOLOv3 - all in Tensorflow.
You'll witness live simulations of Autonomous driving images.
Prerequisites:
Python basics
Basic understanding of CNN architectures (ResNet)
If you are passionate about learning new technologies and methods in tech, this talk is for you.
Laisha Wadhwa
Head of Devrel @Pesto Tech | Ex @Goldman Sachs | Podcast Host| WTM Ambassador| Tech Speaker| Mentor
Delhi, India
Links
Actions
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