Buyside vs sellside quant
If you are weighing a quant offer from a bank against one from a fund, the job titles will not help you much. Researcher, trader, developer. Both sides use the same three words. What changes is the reason the job exists.
A bank earns its money from clients. It quotes them prices, builds products for them, and hedges most of the risk so it can do the same again tomorrow. A fund has no clients paying it a spread. It only earns when the positions it chooses to hold pay off. So the bank quant is mostly protecting the business, and the fund quant is mostly trying to grow it.
A quick note on where this comes from. My whole career has been on the buyside, and I have never worked at a bank. Sellside view in this piece comes from long conversations with peers who do the job there, where quants are often called strategists, or strats for short.
Quant Researcher
Same maths, different purpose. At a bank, a researcher spends most of the week making sure the desk can price what it sells, things like options, swaps and structured notes. The model has to match the market and hold up when the model review team starts asking hard questions. At a fund, the researcher is hunting for signals in very noisy data, testing them hard, and throwing most of them away. Weeks of work can end with the answer that nothing is there, and that is a normal part of the job.
| Day to day | Sellside | Buyside |
|---|---|---|
| Morning | Look at overnight pricing runs and answer the desk's questions about odd numbers | Look at yesterday's live results and check nothing has drifted |
| Core work | Build and calibrate pricing models, produce risk numbers for the book | Find signals, build risk and cost models, decide how the portfolio is put together |
| Who you talk to most | Traders, structurers, model validation | The PM, other researchers, data engineers |
| A good week | A new product gets priced and approved for trading | An idea passes every test and gets a small live allocation |
| A bad week | A model breaks during a market move nobody planned for | Three months of work turns out to be noise |
| How you are judged | Accuracy, stability, passing review | Money made in live trading after costs |
On the sellside, the big banks such as Goldman Sachs, JPMorgan, Morgan Stanley, Barclays, BNP Paribas and UBS, usually in teams called Strats or Quant Analytics. On the buyside, systematic funds such as Citadel, Two Sigma, D.E. Shaw, Renaissance, AQR, Man Group and Qube, along with multi manager funds like Millennium and Point72.
Quant Trader
This is the seat closest to the P&L. On the sellside, the flow comes to you. A client wants to sell a block of bonds or buy some FX options, you give them a price, and then you own that risk until you hedge it or trade out of it. On the buyside, nothing lands on your desk unless you choose it. You run strategies, decide how much capital each one gets, and decide when to pull it back.
| Day to day | Sellside | Buyside |
|---|---|---|
| Where trades come from | Client requests | Your own strategies |
| Main question | What price do I show, and how do I lay off the risk | How much capital should this strategy have right now |
| Typical day | Quote clients, manage inventory, hedge as the book moves | Watch strategies run, check fills and costs, adjust sizes |
| Pressure point | A large client order arrives at a bad moment | A strategy loses for two weeks and you have to work out why |
| How you are judged | Spread earned while staying inside risk limits | Return after costs compared with the risk taken |
On the sellside, the electronic trading desks in FX, rates and equities at banks like JPMorgan, Goldman Sachs, Deutsche Bank and HSBC. Market makers such as Jane Street, Optiver, IMC and Susquehanna sit somewhere in between and hire a lot of traders. On the buyside, multi manager funds like Millennium, Point72, Balyasny and ExodusPoint, and systematic funds like Two Sigma and Qube.
Quant Developer
Every model and every strategy above runs on code someone has to build and keep alive. At a bank, that means a pricing library and risk engine the whole trading floor and the regulators rely on. Getting it right matters more than getting it fast. At a fund, the developer's real product is speed. If a researcher can run a test in ten minutes instead of a day, the team learns ten times as fast, and you see that in the returns.
| Day to day | Sellside | Buyside |
|---|---|---|
| Top priority | Correct, stable and easy to audit | Fast research, fast and cheap execution |
| Typical work | Add products to the pricing library, fix overnight risk runs, support regulatory requests | Clean and load new data, speed up backtests, build trading infrastructure |
| Usual stack | C++, Java, C#, Python, a lot of older code | Python, C++, some Rust, large compute clusters |
| Pace | Careful releases with plenty of testing and sign off | Small teams that ship often |
| Where it can lead | Strats or quant research | Often into research or owning a strategy |
On the sellside, the quant library and strats engineering teams at banks like Goldman Sachs, JPMorgan, Morgan Stanley and Bank of America. On the buyside, funds such as Two Sigma, D.E. Shaw and Man Group, the engineering groups inside multi manager funds, and high frequency firms like Hudson River Trading, Jump Trading and XTX Markets.
Which side to pick
If you like building something exact and solid, the sellside is a great place to start, and you will understand products better. If you want your work measured in money made, and you can live with results that move around a lot, the buyside is the better fit.
On the traditional path, lots of people spend a few years on the sellside before moving to the buyside.