It has been insanely fun working with the AI to build new strategies. You can iterate and test ideas so quickly on QuantConnect now its mind blowing!
I always bake in diversification (uncorrelated assets), asset-level risk control, inverse-volatility position sizing. With those three you can use virtually any universe of stocks and it has excellent performance.
Credit to the Triton Quantitative Trading team at @UCSanDiego for their latest QuantConnect research: a volatility-targeted QQQ–TLT rotation using a multi-scale CNN-LSTM model to forecast turbulence and adjust exposure.
https://t.co/Ddfy9VWSyU
Last night I vibe coded a cookie trading game for career day for 1st graders. We used poker chips for cookies. I think it was a hit -- code here for other dads in finance 😂 https://t.co/9vffHpuBp8
Systematic research is becoming fully agentic.
Today we launched QuantConnect Assistants, an @openclaw like framework for quant.
Specialized AI agents can research ideas, run notebooks, analyze backtests, read logs, and coordinate through a visible Research Pipeline:
Ideas → Research → Validation → Backtesting → Live Monitoring
You can build your own agents, and automate your specific research processes. Read more:
https://t.co/sWGuMi1TxZ
The team at @QuantConnect recently created an integration that brings Databento's real-time and historical market data directly into LEAN, their open-source algorithmic trading engine. For the hundreds of thousands of developers already building with LEAN, adding a Databento API key provides our normalized market data across research, backtesting, and live trading - all within their existing setup.
⬇️ To get started, check out the guide below
Final count down to the Quaint Quant Conference in Dallas this Friday! Our founder @JaredBroad is speaking on Applied AI for Quant Finance ✨ - come say hello!
https://t.co/X2gw4rBCCl
@SystematicPeter Yes QC trades with the spread, slippage and fees. Fees are substantial with this many orders.
The code is attached in the link, I think I implemented your post accurately.
Defi crypto trading with @dYdX exchange is now live!
Easily load historical data in notebooks for research, do accurate point in time backtesting, and deploy your strategies to live-meme-trading in minutes 💪🙈🚀
Time is your most valuable asset, focus on your alpha, not infrastructure.
https://t.co/GVbLI9f9ab
Happy Thanksgiving Quants! We're grateful for you!
This week there is 20-30% off plans, Agentic Quant AI Tokens, and some data on QuantConnect. Check it out in comment below.
We've quietly released Mia V2 to public beta. Its pretty mind blowing 🤯🎉, full agentic quant LLM.
@gabriel1 "Homework for Life" from book "Stories Sell" by Matt Dicks. TLDR: Writing down a tiny note each day helps you remember the details and makes days feel full.
This to me is one of the most important images from the Apollo missions, and its something most people outside space circles don't know about, and may not get the significance of.
It shows Pete Conrad standing next to Surveyor 3, an unmanned spacecraft that landed on the Moon in 1967, and in the background can be seen the Apollo 12 lunar module he arrived in in late 1969. One of the objectives of this mission was to demonstrate high accuracy landing - and they managed to put the LM down a short walk from this probe. They recovered material samples from it to study how it had held up after 2 years on the surface.
What this meant was that, in 1969, NASA had the capability to do lunar surface rendezvous missions - i.e. sending multiple payloads to the same site - up to and including other lunar modules. An unmanned LM that removed its ascent engine and propellant could deliver an extra 3 tonnes of cargo - a manned mission landing near to it later would then have that available for a much longer duration mission. NASA indeed began planning this, with 14 day surface time, but the budget was cut before even the first lunar landing.
So this demonstration undermines claims that Apollo was a dead end. It wasn't - the decision to not continue after it was purely political.