#Jane Street vs Databricks ML

1 messages · Page 1 of 1 (latest)

dull tusk
#

he is on databricks ML compilers team

#

js no idea

worn crystal
#

bruh

dull tusk
#

also insane exits

worn crystal
jagged heron
#

databricks lol

dull tusk
#

Jane Street vs Databricks ML

#

future growth

#

js is only 450 recurring

worn crystal
dull tusk
#

bc databricks insane growth potential better work

#

he will start in august

#

FT

worn crystal
#

yea genuinely pmo

dull tusk
#

He is grad this may

#

what

#

this is for my friend

south sluice
#

what is jane street

loud saddle
#

databricks is mid js>>

clever bane
tight scaffold
dull tusk
#

databricks will probably be worth around 600 billion

tight scaffold
#

that's all smoke and mirrors

#

it could be worth that much or it could pull a snowflake

#

point is ev of js is higher

#

if he cares about money more

#

if he wants to work in tech then db

dull tusk
#

he wants to work in tech long term

#

i do think js ev is higher tho

tight scaffold
#

but it's dumb to choose DB for the money

ocean tiger
#

bro has too much time on his hand

dull tusk
#

@native gyro

clever bane
#

yo he should renege both

#

and work at c1

#

this is a very trivial question

dull tusk
#

290

clever bane
#

swe is like 250

dull tusk
#

fy

clever bane
#

probably like 280 or smth

dull tusk
#

or 280

clever bane
#

makes sense

dull tusk
#

no idea

clever bane
#

wait yeah db is never ipoing

#

ur equity is worthless

dull tusk
#

yeah i think he is going to go js

#

after this

clever bane
#

?

#

wdym

#

jane street will let u defer ur return offer

#

but u need to get the return offer lol

dull tusk
#

it is 73% ro

#

for js

clever bane
#

js ro is largely culture fit afaik

dull tusk
#

no clfif

#

cliff

#

wdym lowered tc

#

ik its 40/30/20/10 vesting percents

#

is grant not same size?

#

damn

#

wtf

#

they sold 😭

#

zon mgso

#

mogs

elder pewter
#

i have yet to see offer where frontloaded but tc is reduced

#

also @dull tusk if your friend db offer already they locked into equal vesting so doesnt matter

#

either but levels fyi says grant size is moving to front loaded

#

its for 2027 tho all ppl ik who got ng offer got equal vesting

#

im ngl ml compilers at db is insane ev

#

can jump to tech labs or quant most likely

#

ik someone who got offer late jan still have equal vesting

#

id assume they doin it for ppl starting 2027

#

ya they removed cliff so u dont need to wait a year to sell

#

imo that signals they gonna be ipoing soon

echo knot
elder pewter
#

true

frail pond
#

😭

frail pond
dawn seal
#

Js is overrated

#

Most people max out at 900k tc with 5+ yoe

#

and they’re working 45 hrs minimum

#

more likely 50-55

#

A swe who would be middle of the pack at js requires 50 hrs but you can be top performer at db / in tech with equivalent 30-35 hrs because the expectations are much lower

#

Rush e6/L6 and then coast 35 hrs/wk with stock appreciation, making 700-1M+

Fuck even ic5 at Roblox makes 700k….

#

don’t need much more atp anyways so u might as well maximize $/hr worked, qol outside wok, and relationships

#

Go db @dull tusk

dawn seal
#

I’ve done the math

#

if you hit IC4 by y5 at Roblox, you make $480k and if ic5 by y6/7, then 590k

#

someone at js could easily do that by working 35-40 hours presumably

sudden horizon
#

I dont think the hours necessarily work out that way

dawn seal
#

DB exit opps are also better than js

#

on avg

#

you’re basically unemployable after 5 years of ocaml

dawn seal
#

ive done the math

worn crystal
sudden horizon
#

lowk ive done the math as well and @dawn seal check outs

dawn seal
#

Jane expects the same output in 40 hrs that google/db/bloc would in 70

sudden horizon
dawn seal
#

no

#

I’m serious

#

Js is 45 min

#

9-6

#

Locked in

#

Others r "40"

#

35 hr actual work

frail pond
#

luck is actually based

frail pond
elder pewter
#

genuinely ive asked 5 ft people + interviewers and all of them said they work 45 hrs

jagged heron
echo knot
#

also you're never even internally promoting to L5 at db let alone L6 😂

#

and then on top of all that you're counting ridiculous amounts of stock appreciation in TC

#

this is genuinely the goofiest comparison oat

#

yeah lmfao NOBODY at db is working 35 hours period

dawn seal
#

I’m just rage baiting tho

#

obviously go js here

#

Y1 IC1:

150,000 + 157,000 * .45 + 20,000  = 240,650

After tax+401k = 144k

—

Y2 IC2:

182,000 + 157,000 * .35 + 125,000 * .33  = 278,200

After tax+401k = 165k

—

Y3 IC3:

220,000 + 157,000 * .20 + 125,000 * .33 + 191,000 * .33  = 355,680

After tax+401k = 206k

——

Y4 IC3:

220,000 + 125,000 * .33 + 191,000 * .33 + 191,000 * .33  = 387,310

After tax+401k = 224k

——-

Y5 IC4:

276,000 + 191,000 * .33 + 191,000 * .33 + 226,000 * .33  = 476,640

After tax+401k = 310k

#

this is the rough math tho assuming u can grind out 45-50 hrs a wk and promo in the first few yrs, comp stays at ~500k after y5 and if you hit ic5 on y8 it jumps to 590k and then normalizes to 700k in abt y9/y10. This is all with 0 stock appreciation, but it’s Roblox so db is def more shit

echo knot
#

Going from IC1 -> IC4 in 5 years is a pipe dream

#

might asw say you're gonna go to js and be one of the swes that makes 3m+/year

elder pewter
#

isnt l2 at db like 400k

#

second row has 500k tc after working 2 years at db

Screenshot_2026-03-03_at_10.09.49_AM.png
#

im just looking at where yoe == years at company

pallid escarp
#

would only be a real discussion if it was mlr tbh

lilac tulip
#

this is rtarded

#

jane street mogs

echo knot
flat matrix
#

i havent someone point this out yet, probably because no one here really knows what ML compilers at dbx is, but its not a high impact job as one might think. the company is a great place for ML sys especially if youre in mosaic or ML infra org because they are making a lot of strides in that field (namely in data but still). See: MLflow, MLlib, clipper, ray, yggdrisil.

However, they arent doing a lot of cutting edge in ML compilers afaik. its mainly using other tools (like TVM) and weaving them into their inference engines and spark library. Good companies for ML compilers where actual innovations are being done would be: Torch compilers at Meta, Triton team at OpenAI, whatever team at Anthropic, Neuron or Annapurna at AWS, CoreML team at Apple, JAX/XLA team (under CoreML org) at Google, Tesla AP, TRT team at Nvidia, modal, modular, and a few other startups.

upbeat condor
#

if its not in hand now u cant reasonably say it will be in hand at some point

#

and anyways whats the chance you stay for 4-5+ years?

#

or not even 4-5 yeras, 2-3

upbeat condor
#

dont think the money should be the decision maker here

covert birch
#

db glaze goes crazy

upbeat condor
#

anecdotally a friend fy 700+

#

to be fair he had similar competing

tardy elbow
#

Why in gods name would you take db over js

#

Things alrdy valued at 100B

cinder thunder
jagged heron
#

js is some ass 😂

cinder thunder
#

gng ur getting slimed out

jagged heron
#

databricks ng mogs ceo of nvidia

cinder thunder
#

Purdue got this mf movin different

jagged heron
cinder thunder
jagged heron
#

that will make me 10 million when databricks IPOs

cinder thunder
#

Robinhood not gonna list them making too much money from rvi

jagged heron