This post shows how to heap dump a running Python process using pyrasite and guppy3.

This post lists materials that I find useful during my journey of learning neural networks from scratch.

f

This book shows how a team can execute an important goal successfully in the face of massive amount of day-to-day operations that are urgently required to keep the business running.

What’s a good design? How to produce a good design? This book provides some answers.

For B2B high-tech startup, there is a period when the product is already successful in the early market and is trying to enter the mainstream market. This period is called crossing-the-chasm period and it’s peril. Many startups die during this period and never reach the mainstream market. This book tells us why there is such a period and why it’s dangerous. It also shows how we can get through this period and be successful in the mainstream market.

When you ask people if they want to be multipliers who make surrounding people better or diminishers who make everyone worse, most people will say that they want to be multipliers. However, in reality, many people are accidental diminishers in some aspects. Being a true multiplier requires a multiplier mindset and certain approaches. This book tells you what multipliers are and how to become one.

I had worked for the LinkedIn graph database team for 5+ years and we successfully built a graph database serving the entire LinkedIn economic graph. In this post, I want to share what I have learned. Disclaimer: many of the words and wisdom are from my great colleagues.

As engineers, our job is to solve problems. In order to do that, we need to discover them first. To me, the best way to find problems is through doing things while keeping the eyes open.

For the discussion of this post, response time is the time between a service receiving a request and returning a response. It is the sum of waiting time and processing time. Waiting time is how long the request waits in queues before being processed. Processing time is the time to actually do the work of the request. Throughput is the number of requests that are completed per unit time. This post discusses how they can be possibly related.