MemSQL is a Y-Combinator, SF-based startup that is leading the next wave of big data. MemSQL’s real-time database is the fastest performing and most scalable database for transactions & analytics. MemSQL is built on a distributed system with a lock-free data structure able to ingest limitless amounts of data with zero downtime. Features include geospatial functionality, connectors & integrations allowing companies to combine semi-structured data all in one place. Those early to take advantage include customers: Comcast, Boeing, Samsung, Zynga, etc.; Investors: NEA, Accel Partners, Khosla Ventures, In-Q-Tel; capitalizing MemSQL with $50M across just two tranches of investment; and we’ve hardly scratched the surface.
Many of our original engineers are Topcoder Open Algorithm finalists and prominent IOI and ICPC medalists. If you have the track-record to be part of this elite team, we want to talk to you!
There’s a reason why MemSQL attracts and hires many topcoder members. Building a database is an extremely hard task. We write very complex C++ code that needs to be fast and bulletproof. We cannot hire average engineers to do that – we need people who are insanely smart. Incidentally, this is exactly the quality you need to become a good Topcoder. Clearly, there are many skills we need at MemSQL that topcoder members might not have. We’ve seen that smart coders can quickly learn these skills, but average coders, regardless of experience, cannot learn to be smart.
There’s a reason why Topcoder members like working at MemSQL. First of all, MemSQL tackles some of the hardest problems on the market. Working with the top engineers in the industry pushes the bounds of what you think is possible. You will see technologies and ideas that have never been built before. But since almost every company in the world claims that the problems they solve are the hardest, we will be more specific:
1. Our entire storage engine is lock-free. Building lock-free data structures and algorithms is very challenging. So is testing it. As part of building the storage engine we found fundamental bugs and errors in several books and whitepapers that are well-known and loved in the industry. Also, check out our post about common lockfree pitfalls in C++.
2. MemSQL is a distributed database. Managing the cluster is a hard task. Servers die in random order and quantities and without any warning, and the cluster needs to continue operating without losing data.
3. Building a SQL query optimizer is far from trivial. People who use databases write insane queries, expecting that the query optimizer will figure out what they actually mean and how to execute that query with decent performance.
Some other things we work on that have interesting subproblems include column store, database replication, and JIT translation of SQL queries into machine code. Most important, all the areas listed above have open work items, so if any of them sound interesting, chances are we need someone to work on them. Not to mention you will be working in our new SOMA office where we have catered meals, snacks, monthly team events, office dogs, and ping pong!
The biggest problem Tocoders face when they start working full time is that their full skill set is severely underutilized, minimizing their impact. Layers of management and bureaucracy have wasted the talent of many red topcoder members already (ask your more senior friend what they have been working on in the last 6 months). MemSQL today has the biggest commercial database cluster ever deployed in the world and is written by a relatively small team. You will have a huge impact at MemSQL, which is only bounded by your skills and desire to learn and code. Send your resume to firstname.lastname@example.org.