lmeyerov

62 ms·

661.
by lmeyerov 4y ago
Here's an attempt at quantifying Junior staff need constant help, but senior staff need space that's hard to achieve at an office. Studies show people take 30m-1.5hr to dial in for deep work. So, imagine a "light" office
662.
by lmeyerov 4y ago
Yep! We do the switch to parquet, and then as they say, use dask so we can stick with python for interesting bits as SQL is relatively anti-productive there Interestingly, most of the dask can actually be dask_cudf and cudf nowadays: dask&#
663.
by lmeyerov 4y ago
Australia (Canberra) for this one, half-remote
664.
by lmeyerov 4y ago
One concrete example we're hiring in: We're looking for a cleared security / SIEM engineer in Australia to help build out GPU/AI SOC tech. As long as they love Python and modern bits like scaling detection engineering &a
665.
by lmeyerov 4y ago
Yep traditional advice is don't burn > 100k / mo when early stage, and most lower. So 2mo+ payroll fine for a normal seed, early A. Later stage are normally more sensitive. It's also a real issue with profitable companies
666.
by lmeyerov 4y ago
It sounds like a test harness bug if startup time matters, the setup here is quite suspicious JSC vs V8 is interesting... I'd like to see something like tsc compile times for real projects compared. And then again against something lik
667.
by lmeyerov 4y ago
Get professional help ASAP vs asking unqualified techies Your employer is likely down for you to go part-time. Not clear if unemployment or new job will actually help, and even if so, way better to find a new job from a position of employme
668.
by lmeyerov 4y ago
The comparison would be renting, not buying the GPUs I believe capex <> opex is more 1:1 nowadays, so something feels off here...
669.
by lmeyerov 4y ago
Slight tangent inspired by the premise, but why do EKS etc cap the number of pods and other resources on a node? We try to enable our customers to start on 1 node and scale up at their discretion, but found we had to into all sorts of conto
670.
by lmeyerov 4y ago
Great writeup by the Kuzu team! An especially interesting area to us is how to redo these to be accelerated with GPUs. Ex: See https://www.adms-conf.org/2014/adms14_wu.pdf We've been taking a worse-is-better appro
671.
by lmeyerov 4y ago
The "no" is bc challengeable reasons, which are useful to be explicit about "This would take resources away from delivering X and limit devs ability to collaborate, so we need a way to motivate the budget precedence / gr
672.
by lmeyerov 4y ago
- Incorporate. In theory, if still an issue, get a bit of liability insurance and whatever else to boost your internal + external reliability ratings... - .... But I'm guessing that's not the problem. It can be you and they are be
673.
by lmeyerov 4y ago
They don't have to. Ad & search dependent companies need to answer for themselves how to handle the coming disruptions. As an analogy, Kodak was a disruption opportunity, not a problem, for apple and flickr. Yes, maybe some content
674.
by lmeyerov 4y ago
Yes. This is a variant of building an OSS project using company funds then leaving the company to commercialize it. It's a real path, but requires taking advantage of years of corporate largess. Not impossible, but some big "if&qu
675.
by lmeyerov 4y ago
Graphistry | Security data scientist, Senior security Python/SIEM solutions engineer, Lead platform engineer (remote) | Hybrid remote - Australia + US | NV1/2 Clearance sponsored | Above-market pay Graphistry is the first visual g
676.
by lmeyerov 4y ago
Senior high-salary hiring is definitely happening at high numbers, but specializing and finding the right opening matters. We get a ton of CVs, but few in our advertised areas, and we know we can keep looking so everyone involved wins. If y
677.
by lmeyerov 4y ago
We have 2 senior openings right now, and both feel representative in not requiring a PhD. We're pretty cutting edge here (end-to-end GPU acceleration, graph neural networks, win R&D competitions, ...), and our team is split pretty
678.
by lmeyerov 4y ago
Great question! For context, I transitioned from publishing top academic papers in security to building & growing a visual graph AI startup where, for one of our bigger customer bases, we work with top enterprise & military security
679.
by lmeyerov 4y ago
Internally at Microsoft, the flops were obvious. Employees made fun of the zune, there was a lot of embarrassment that staff were preferring iphones to free company-provided alternatives, etc. Leadership was putting $ in (remember the Nokia
680.
by lmeyerov 4y ago
Use numba and it is, including on GPU :)
681.
by lmeyerov 4y ago
(not them, but in this space with major enterprises and gov agencies deploying Graphistry) We are pretty active here with security cloud/on-prem data lakes teams as a way to augment their Splunk with something more affordable & res
682.
by lmeyerov 4y ago
Totally! My point is chatgpt isn't a high-moat advantage for text/q&a for microsoft. Their top competitors here have similarly huge UI footprint. In contrast, program synthesis has a much higher data moat.
683.
by lmeyerov 4y ago
Yes & No. GPT <> ChatGPT: probably not. It's not hard for other big players to enter this space. It's mostly egg-on-face for Google that they haven't given that Google basically invented the model that OpenAI uses a
684.
by lmeyerov 4y ago
That's... a lot of specialized folks To help make lemons into lemonade, we're hiring for folks into webgl2 for the next gen of our GPU client <> GPU cloud visual analytics engine used by analysts to investigate problems like
685.
by lmeyerov 4y ago
For Graphistry's GPU UI, we solve via react code for HTML chrome & interactions, and the visual analytics in webgl controlled by rxjs. Beyond rxjs being an overall coordination language, it lets you do thinks like custom schedulers
686.
by lmeyerov 4y ago
Yep we use cuGraph a ton, and stick to the CPU stuff for < 100k graphs, esp for their more niche algs. Exciting days!
687.
by lmeyerov 4y ago
Yep. Our biggest deals (7 figs) are remote. I do buy into getting on a plane or having someone local. That generally speeds up and increases close rate + renewal. But if you nail distribution or PLG.. can run enterprise sales much more effi
688.
by lmeyerov 4y ago
Probably for SQL (top n, ...), but not for wrangling & analytics & ML & ai & viz
689.
by lmeyerov 4y ago
We're quite interested in a related case here of byte-addressable CDN caching of data table reads for web-scale publishing. So right in that gray area of web publishing. Imagine something like an infinite on-demand parquet table scroll
690.
by lmeyerov 4y ago
Graphistry | 3 openings - Platform engineer, security data science, security solutions engineer / security engineer | Full-Time or Contract | DC, US, Australia, Global | Remote + Hybrid | $125K-225K + equity + profit-sharing | https:

More ›