Morgan Stanley quant MD: AI-assisted Python coding is a "revelation"

15 October 2025

3 minute read

Morgan Stanley is one of many banks to implement AI coding tools; it famously developed a legacy code automation tool, Devgen.ai, that is thought to save over ~200,000 hours of work for its engineers. It's not just engineers benefitting from AI coding, however.

Click here to register your interest in entering the bubble by eFinancialCareers, our new anonymous community. ✍️

Speaking at Quant Strats Europe today, the bank's global head of research for quantitative investment strategies (QIS), Stephan Kessler, said he and other Morgan Stanley quants are using coding tools; "if you're using Python and you don't use GitHub Copilot... [implementing it is] a revelation." Kessler says AI tools are "speeding up our coding process by multiples," allowing quants to code more complex things much faster.

Morgan Stanley, of course, isn't the only bank using AI. Bank of America MD David Shelton, also speaking at the conference, said that the tools are becoming increasingly popular at his bank, and that they significantly speed up the ideation process.

Other conference panelists beyond banking were also espousing the benefits. Alexandra Bilon, global head of Blackrock's fixed income portfolio analytics group, said that Python "is our favourite language at Blackrock." She said that using tools like CoPilot and windsurf is "amazing" for improving how fast you can code. She said it also improves "how fast you fail," so you don't have to waste time on subpar ideas.

Hamza Chaudhry, a fixed income quant researcher at investment manager Alliance Bernstein, said that "any researcher that isn't using [AI coding tools] is going to fall behind." He said AI is particularly useful in a debugging context. For code that would have taken Chaudhry hours to fix, "the latest version of Claude can immediately figure out where the bug is... Claude is super cool."

Morgan Stanley doesn't just use AI for coding, though. Kessler said a key use case of AI for quants is signal generations from "sentiment analytics." by analyzing news reports (hello Morgan Stanley's AI). He says AI is also particularly useful in the bond market for analysing and compressing the information of bond prospectuses. Kessler said quants "get emailed hundreds of them daily... it took days for an analyst to do one bond, now we can do thousands of them in minutes."

Kessler advocates for quants to "form a symbiotic relationship" with AI, and to "rewire our thought processes" to make us more able to use the tools. If you don't, the future may be rough.

Have a confidential story, tip, or comment you’d like to share? Contact: WhatsApp: http://wa.me/442079977910 (+44 20 7997 7910), Telegram: @AlexMcMurray, Signal: @AlexMcMurrayEFC.88 Click here to fill in our anonymous form, or email editortips@efinancialcareers.com.

Bear with us if you leave a comment at the bottom of this article: comments are moderated intermittently by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. You must take sole responsibility for comments you post on this site. We will take reasonable steps to weed out anything that we consider to be offensive or inappropriate.

Photo by Tsuyoshi Kozu on Unsplash

AUTHORAlex McMurray Reporter

QX
QXA
22 October 2025
Python is an extremely powerful tool (like AI) in the right hands.... in the wrong hands it can be the destroyer of your company. Unfortunately, the management prefer to hire low quality dev with the AI tool in mind, thinking it's easy and we can do cost cutting. The reality is python with AI can create technical debts at a fastest pace never seen before. But only senior technical people can understand that point. For the business : Great, I can write whatever I want in a record time. They will be stuck in a endless support loop. output : They won't see it because they are focus on something else. But it will cost them more money and time. For the junior dev : Great, I don't need to learn deep concept, my AI friend can do it for me. Lead to poor quality code, no design, bugs in prod and hell of maintenance. output: Lower pay, no evolution, lost of knowledge. For MDs : Great, we can do some cost cutting and hire an army juniors. They will received a big bonus as reward of the cost cutting, the fast delivery of projects and create a endless loop of bugs & fixes in prod (auto generation of jobs). The Business will have the fake impressions that the team is responsive and won't realized that the issues come from the poor quality. output: for MDs is relatively positive, more money, more budgets, more people, better image. For senior technical people : Risk to be fired and replaced by low quality dev people. Long nights trying to fix the poor code quality and design. They will eventually leave the company to create their own solutions/project in smaller company where the quality is the main driver. Small company can not absorb the "fat" like big banks or hedge funds. output : pretty bad on almost everything in the company. Really good outside the company, they will be able to produce more and create solutions faster. The Future: 1- CTOs and CEO will be focus (already are) on cost cutting, will fire the seniors and replace them my army of juniors with AI. 2- The quality of the software will drop drastically. More bugs will appear. More supports required. More people needed. 3- The only jobs for juniors will be low dev with AI. Meaning we are going to lose knowledge on how to do good design and good code. 4- No promotion anymore, junior will just use AI, no need for them to learn more, the AI learn for them, meaning they are expendable. 5- Some CTOs and CEOs will call back the seniors to bring back quality but it will be too late. 6- The AIs across the world will learn only from code written by junior using AI. Garbage in , garbage out, the quality will be even lower with time The AI with python used without sense will cost a lot of money, time, job and career, . With a little output. High price to go down ... but I'm sure we will see
Reply
Di
Disgruntle
16 October 2025
Yes, financial professionals who only learned to code out of necessity doing trivial things are sped up by AI. Ask any experienced developer and you get a different story.
Reply
Ho
Hopeless
16 October 2025
Very interesting and definitely the way to go.
Reply