Here's the #1 thing I see people miss when building a trading platform or backtesting engine:

A great trading platform nails ops issues - tradeops/sysops/devops/MLops-style problems. It's so important to get these right. 🧵 Things like how you manage model configs and versions, generate meta configs, manage multiple instances, ship logs, configure sessions/ports, configure kernel & network params, handle crashes & persistence, interact with other applications, interact with multiple gateways/brokers, pick up intraday fixes, deploy, sync post-trade, etc.

Thread incoming: Why China will win the AI race in 8-10 years (and why that might be good for the US) 🧵

The core reason: Chinese people are willing to trade privacy for convenience. Americans are paranoid about their data being sold and will go out of their way to make sure that their data isn't being shared. This gives China a massive domestic market advantage - imagine training AI systems on a population that's comfortable sharing data at scale. It's like having a home field advantage in the biggest stadium in the world.

My DeepMind friends think it'll be sooner than 8-10 years.

IEX Cloud is @DatabentoHQ's competitor, but my heart sank upon their shutdown. Here's my takeaways:
1. Sales > Product. It's about who has the best sales team, not about who has the best product (for better or worse). IEX Cloud relied purely on PLG, with 0 SLG. 🧵 2. Institutional >>> Retail. Selling to retail customers (or even retail platforms) isn't profitable.

"What do you do that AI can't do?" I worked in HFT. Below are the limitations of AI in my industry. 🧵 While AI has made enormous strides over the last decade, there are 8 reasons why AI hasn't consistently predicted the stock market. (It's like 8 floors of a dungeon that AI has to defeat before winning the game.)

Here's a hard truth that quant firms can't say out loud: According to multiple sources from Tier 1 quant/HFT/prop shops, the number one indicator of success as a quant trader is the university they attended... for undergrad. 🧵 People will always disagree, and there are plenty of exceptions at Tier 1 firms. However, these are some of the best data scientists in the world. I'm inclined to believe them, because their livelihood depends on their ability to wrangle data and remove biases.

"Buy vs. build" is a classic problem many startups face. For example, on the hedge fund technology side (think: OMS, data feeds, market simulation, etc) there's some classic tradeoffs and risks to consider: - Imagine you're looking for software or data, but nothing fits your criteria. Existing solutions match 70% of what you need, but there's frustrating gaps. So you consider building it yourself.
- That's what we did at Domeyard, but we ended up suffering some consequences:

(On HFT data streaming) The challenge is to build servers that provide a simple firehose AND more stateful capability to manage customized subscriptions and filter symbols. Here's how we did it: 1. Keep things simple and choose boring technology.
The funny thing is, coming from the HFT market making world, we weren't very familiar with the modern tech stack & frameworks for distributed streaming. In HFT, distributed processing is often an antipattern.

#quant "What separates the top 3-5 firms from the tier 2 firms after them?" - a thread by my colleague I've carried out a lot of #quant research interviews & ran a fund. I've noticed that junior candidates tend to obsess over vanity metrics like alpha, latency, AUM, Sharpe, etc. These are important, but here's 3 things that are probably more predictive of a firm's success.

Andrew Chen (Partner, Andreessen Horowitz + Twitter influencer) and Emma Waldron (former Miss Ireland) lied to illegally obtain a permit to get married on protected land, trashed the place, destroyed vegetation, and suffered 0 consequences.

Deleted video + evidence below: @30minly got the receipts:

Most people should stop writing RESTful APIs for financial data. - A thread by a colleague who doesn't use Twitter 🧵 I think this trend is driven by three reasons:
1. Many newer fintech firms were inspired by the crypto exchange wave and just copied these crypto exchanges' REST APIs blindly, without having seen best practices in mature asset classes.

“Risk is randomness that you can parameterize. Uncertainty is randomness that you cannot. The best opportunities come from taking on uncertainty, not risk.” - Andrew Lo In any data-heavy field, you'll hear the terms "risk" & "uncertainty." Examples of uncertainty in finance is the occurrence of unexpected events or market shocks. COVID-19, wars, regulatory changes, and other events can have a profound impact on the markets & investment returns.

#HFT strategy time, FINAL part 5: Optimization. See comments to catch up on this series.

--
We can try to optimize values of, say, 1 ≤ k ≤ 5 and 2 ≤ L ≤ 6 differently for each market in steps of 0.5 for every symbol. 💡 General tip: Try to keep the search grid small within your domain knowledge as: (1) the more the number of parameter sets that you try, the higher the risk of overfitting, (2) large scale optimization can be computationally slow.

#HFT strategy time, part 4: Strategy implementation. See comments to catch up on this series. Why skew for this example? 
➡️ It uses order book data. Many (but not all) HFT strategies use order book data b/c it’s almost like a third dimension, revealing more about the markets than top-of-book data. 
➡️ It’s applicable to both low- and high- frequency strategies.

#HFT strategy time, part 3: Strategy parameters and considerations. 🧵
Continuing my HFT / quant example, context and disclaimers afterwards. A simple rule-based strategy is to buy a minimum quantity (1 lot) when the skew is highly positive and sell a minimum quantity when the skew is highly negative. Since this should be symmetric, we can introduce a single parameter k to decide when we’d buy or sell:

#HFT strategy time, part 2: Examples of features.
- Skew: The simplest type of book feature is called the skew, which is the imbalance between resting bid depth (v₂) and resting ask depth (v₁) at the top of the book. So we can formulate this as some sort of difference between v₂ and v₁. It’s convenient to scale these by their order of magnitude, so we take their log differences instead:
skew = log(v₂)−log(v₁) = log v₂ / v₁

#HFT Here's an example of an HFT strategy. Keep in mind this is 1 in a billion examples. This isn't investment advice. Don't trade it. I'm in a unique position where I can share strategies without getting sued cuz it's my company & IP. #HFT Part 1: Important terminology. It's hard to jump into an HFT strategy without explaining some terminology first.

The SVB collapse has been devastating in more ways than one: They supported women, minorities, & the LGBTQ community more than any other big bank. This includes not just diverse events, but actual funding. SVB helped us move one step forward; without them, we move two steps back. 2) Many companies had cash in SVB, including companies like Etsy. This impacts independent makers and artists. Even if you didn't bank with SVB, you'll probably start to feel the impact of it over these next few weeks.

BREAKING: SVB has been shut down. “The most recent statement from the FDIC is that all funds up to the insured amount ($250k) will be available in the short term.” fdic.gov/news/press-rel… My last post got so much hate. Some people judged me super hard and lost respect for me. So I’m saying this one more time.

Thank you. I’m grateful to be able to help other startups launch faster with our data. I have enough cash to live my life pretty comfortably. I’m sorry that you have a problem with my personal career choice. Honestly I’m a nobody and will never understand the obsession lol

https://twitter.com/_pablomitchell/status/1634103717665583105

I wasn’t successful by any means. Maybe a little bit (millionaire) but nowhere close to the billionaires you see on CNBC. So the obsession with my every move (treasuries, data, etc) boggles my mind sometimes. Thank you, internet stranger

Idk about you but I make enough money that, as a busy CEO, it’s worth it to pay a tiny fee to get an LAL, improve our credit, and access treasuries. 🤷🏻‍♀️ Outsourcing admin work is awesome.

https://twitter.com/_pablomitchell/status/1634096798133547009

I grew up poor and my parents never taught me the time value of money. They would waste hours doing taxes themselves and translating stuff when they could’ve hired a cheap tax preparer. Likewise, I used to hand-sew my first company’s swag cuz I was THAT cheap.

I was supposed to do an interview this morning, but the candidate ghosted me. After emailing him and waiting 20 minutes I gave up and left the Zoom. Later he emails me being like "sorry can you delete my application." I didn't mind his response, but definitely learned a lesson: 1) It's a red flag when candidates waste an interviewer's time, and vice-versa. If I was the only decision-maker, I would reject everyone who ghosted me, even if they apologized later. Unless you had an emergency, you should reschedule *before* the interview, which takes ~1 min.