Ask HN: Who is hiring? (February 2024)

If the assignment really only takes 30 minutes, why can't it be a live call if the candidate wants?

whether you like it or not, take home assignments just select for the people who can spend the most time on take home assignments.

Which means people like me who have a current job and a family and can't afford to spend several hours on a coding test to make sure they're above and beyond the competition either won't apply, or will send in results that will be much often worse than the competition because we actually only spent 30 minutes on it.

[dead]

Actively | Software Engineers (LLM/backend/data infra, very product-y) & Forward Deployed Engineers (great data/SQL chops, great at understanding customer problems, and strong engineer in general to automate work using LLMs/internal tools) | Full-time | NYC | Onsite

I'm one of the founders of Actively. We're building best-in-class machine learning and LLM-powered scoring for outbound sales teams. Our first product lives in Salesforce, ingests data from a variety of sources (firmographics, intent signals, marketing interactions, and prior engagement, including unstructured call transcripts, email chains, and notes), and combines LLMs with structured models to tell sales reps who they should reach out to each day (and why) to be laser-focused.

Our paying customers are fast-growing unicorns you've definitely heard of with large outbound sales motions, where the reps use us to target their outbound every day. Our thesis is that scoring (a known/"need to have" part of the stack -- but able to be done 10x better with LLMs because you can incorporate the goldmine of unstructured engagement data) is the right wedge into building the intelligence-driven outbound sales stack of the future that _actually works_ for large companies that need to generate millions of dollars in pipeline each quarter. Unlike most other companies in the space, we are explicitly building for companies of this scale and have them using our product today "in production." We have a 100% close rate from pilot → paying customer and want to scale our customer base aggressively this year.

We started with the vision of bringing AI to every business team, and have since narrowed down to focus on outbound sales teams who have a bunch of attractive properties (there are a _ton_ of them out there, they all use the same stack, the data is consolidated in one place, lift is measurable and directly touches the top line, etc.).

We're a small team from Scale/Waymo/Google/etc. (including some very senior folks) working in-person in NYC and looking to bring on a couple more engineers as we start scaling our customer base, in both product engineering & forward deployed engineering roles. If you have at least a few years of experience, are highly entrepreneurial / excited to work closely with customers + dig in across the business, and are excited to work on some combination of LLMs/data infra/across the stack, we would love to chat: mihir [at] actively.ai!

Keywords: AI, ML, LLMs, machine learning, Salesforce, explainable AI, product engineering, SQL, forward deployed engineering, solutions engineering, data infrastructure

i have plenty of experience with c++, python (occasionally with types), linux, postgres, implementing web servers, implementing browser frontends, low-level performance optimization, implementing distributed systems, implementing academic papers, and customer interaction; i've dabbled in typescript, react, and haskell, and last week i was pair-programming with a friend who was kind of porting a raytracer i wrote to cuda. another friend of mine spent the summer doing research on the ceph team, and another one runs nix and ceph at home and has been experimenting with rolling nix out across his company's fleet. i've tried guix myself but not nix

so i don't think this is such a crazy list; their stack is nearly all like super mainstream technology, except haskell, and haskell is hardly some unknown language