Your biggest problems will be brokers, data, and capital. The coding part of quant finance is surprisingly boring (unless you are doing HFT but most aren't). Fo

Your biggest problems will be brokers, data, and capital. The coding part of quant finance is surprisingly boring (unless you are doing HFT but most aren't). For infrastructure just use AWS and co-locate next to the exchange (especially if you are doing crypto where everything's in the cloud). For data try https://databento.com/ https://databento.com/, there are also other sources. Affordable data, especially tick level L3 data at scale is difficult to get. Having enough capital to make trading worth well (pricing in losses etc.) is also difficult. Direct market access etc. all cost money. Use interactive brokers if you need a compromise between fully professional and amateur. Spend more time on your financial models, and less time building beautiful frameworks. The engineering of quant finance is the least interesting part. Focus on the market microstructure and financial models. If your model is bad, you lose money regardless of how great your infra is. It's difficult to find a good off the shelf backtester. This ties into your model too. At its core, a backtester is just a for loop. Computation of variables like slippage and integration of monte Carlo Sims is where the secret sauce lies. Those are all proprietary. Best of luck in seeking alpha