dalianaliu LinkedIn Posts

It's frustrating that the title 'data scientist' can mean so many things.

I've seen ML enthusiast got hired into data engineer roles.
They ended up

In 1999, nobody thinks Amazon can make it.

Today, Amazon has revolutionized both retail and cloud businesses.
"Listen and be open, but don't let

Ludwig, an open source project developed at Uber, makes it easy to rapidly build and deploy state-of-the-art deep learning models in 15 lines of a config

It's not about how much data you have, it's about how much meaning can you unlock -- and that's why we need embeddings.

A/B testing -> where intuition meets the unexpected.

I recently talked to the 4th data scientist at Airbnb on the the unexpected challenges

How to hire your first ML team in 2023

You'll 100% discover new problems while working on ML projects. There hasn’t been any project

When I was building my first computer vision model, it wasn't hard to use transfer learning to get a baseline. However, it

"ChatGPT Code Interpreter will replace data scientists."
--- said by people who never worked in data science.

It's like saying "data science is dead"

Let's talk about data quality for ML systems👇🏼

Join me for Data Quality Day, a half-day virtual session featuring insights from data teams on

I'm interviewing Aishwarya Srinivasan on The Data Scientist Show, what should I ask her??

If you don't want to miss this episode, subscribe here:

•

AI is a data product.

Gone are the days of creating random chatbots just to say you’re building AI. In 2024, teams are likely to become

Data scientists are feeling tired 😩

So, we met in New York, ate pizza, and ranted about our frustrations!
Here are the most common ones,

LLM Developer Meetup @San Francisco:

We brought together founders and engineers from leading LLM startups and open-source projects for one night to share

I had the pleasure of being one of the first ever to interview Daliana Liu on a podcast! We talked about how developer

Building LLMs can be lonely.

Let's talk about LLMOps with food and drinks! I'm going to a meetup in San Francisco next week

Building LLMs can be lonely.

Let's talk about LLMOps with food and drinks! I'm going to a meetup in San Francisco next week

I'm hosting an online fireside chat on how to accelerate your A/B testing.

If you want to get from 10 experiments to 1000, here

Only 10-20% of the A/B tests have positive results at Google and Bing!

This means that companies need to run *many* experiments to find

Last month, I asked 6 data scientists: what makes you tired?

Here are some answers...
• the endless follow-up requests
• manage your managers
• context-switching
• job

Building AI products requires experimentation.

Multi-armed bandits (MAB), with its roots in reinforcement learning, offers a dynamic and efficient approach for balancing exploration and