The rush for expertise in bringing Chat-GPT to market in financial services has begun.
In March, Bloomberg announced the launch of its own Large Language Model (LLM), Bloomberg GPT. It's now begun hiring for that team in its CTO office, with two big roles currently available. One is for a technical product manager who "will have ownership over our technical strategy for large model training and serving." The other is for an applied LLM researcher who will be "working on strategic research and bringing that work to clients through product development."
At the upper end of their salary ranges, both roles are paid extremely well. The product manager can earn a salary of $295k while the researcher can earn $330k. However, comparative to other research roles in that area, they aren't quite as impressive. San Francisco based research firm Anthropic has upped their salary range for a prompt engineer and libarian role we previously highlighted as paying $335k. Now, it can earn $375k. A research engineer with "extensive engineering experience" meanwhile can earn a staggering $450k.
The product management role is all about, 'ownership' of Bloomberg's 'technical strategy for large model training and serving.' The more highly paid research role is about helping "code generation to revolutionize how engineers write code."
The two roles come after Marco Argenti, Goldman Sachs' CEO, said earlier this month that a talent shortage is is emerging for people who "know the latest LLMs and transformers." Goldman too wants LLMs to write code and Argenti said “up to 40%” of the code generated in tests was usable.
Are prompt engineering roles the future?
Prospective computer science students might be thinking of packing away their ambitions of learning how to code. Speaking to eFinancialCareers, Vivek Nashar, CEO of testing company HackerRank said knowing how to code in future will be more about, "putting in the right prompt, interacting the right way, knowing the limitations and knowing when to let it do the work on its own."
It may not be that simple. Bloomberg wants its researcher for have a PhD in "code synthesis, information retrieval, natural language processing, and/or deep learning research."