#258 Machine Learning for Ride Sharing at Lyft, with Rachita Naik, ML Engineer at Lyft

Nov 4, 2024

Rachita Naik, a Machine Learning Engineer at Lyft and a Columbia University grad, dives into the fascinating world of machine learning in ride-sharing. She discusses the complexities of ETA predictions and the real-time algorithms that keep users informed. With an emphasis on dynamic pricing and safety, Rachita addresses the challenges of late model deployment and latency issues faced by millions. She also highlights the innovative role of AI, particularly generative technologies, in enhancing customer interactions and driving continuous improvement at Lyft.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

ETA Prediction

ETA prediction is crucial for ride-sharing apps, using algorithms to estimate arrival times.
These algorithms consider historical ride patterns, real-time traffic, and weather conditions for accurate predictions.

INSIGHT

ETA Pipeline

ETA prediction pipelines involve multiple layers, including mapping, real-time signals, routing, and bias correction.
Lyft uses a system called LyftNav, incorporating driver GPS, OpenStreetMaps, and real-time traffic data.

INSIGHT

Challenges of Scale

Building large-scale ML systems presents challenges like latency, scale, and model degradation.
Lyft prioritizes system reliability and uses fallbacks, shadowing, and testing to ensure smooth deployments.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Machine learning and AI have become essential tools for delivering real-time solutions across industries. However, as these technologies scale, they bring their own set of challenges—complexity, data drift, latency, and the constant fight between innovation and reliability. How can we deploy models that not only enhance user experiences but also keep up with changing demands? And what does it take to ensure that these solutions are built to adapt, perform, and deliver value at scale?

Rachita Naik is a Machine Learning (ML) Engineer at Lyft, Inc., and a recent graduate of Columbia University in New York. With two years of professional experience, Rachita is dedicated to creating impactful software solutions that leverage the power of Artificial Intelligence (AI) to solve real-world problems. At Lyft, Rachita focuses on developing and deploying robust ML models to enhance the ride-hailing industry’s pickup time reliability. She thrives on the challenge of addressing ML use cases at scale in dynamic environments, which has provided her with a deep understanding of practical challenges and the expertise to overcome them. Throughout her academic and professional journey, Rachita has honed a diverse skill set in AI and software engineering and remains eager to learn about new technologies and techniques to improve the quality and effectiveness of her work.

In the episode, Adel and Rachita explore how machine learning is leveraged at Lyft, the primary use-cases of ML in ride-sharing, what goes into an ETA prediction pipeline, the challenges of building large scale ML systems, reinforcement learning for dynamic pricing, key skills for machine learning engineers, future trends across machine learning and generative AI and much more.

Links Mentioned in the Show:

New to DataCamp?

Learn on the go using the DataCamp mobile app
Empower your business with world-class data and AI skills with DataCamp for business