Chicago · AI, ML and LLM engineers
Chicago AI Staffing Splits Into Two Pay Scales With Nothing in Between
The metro benchmark for computer and mathematical work is $52.08 an hour, under the national figure. Almost nobody building models in this city is paid anything close to it. We recruit both ends of the split on contract, contract-to-hire and direct hire.
Chicago AI staffing covers machine learning, LLM and applied AI engineers across the metro, where computer and mathematical work averages $52.08 an hour, below the national figure, and KORE1 returns a qualified shortlist in 17 days. We cover the Loop and Fulton Market, the North Shore and Lake County, the I-90 northwest corridor and the I-88 research belt out to Lemont, on contract, contract-to-hire or direct hire, as part of our wider IT staffing services practice. Twelve-month retention on those placements runs 92%.
Last updated: September 1, 2026
$52.08/hr
What the whole computer and mathematical group averages across Chicago-Naperville-Elgin, per the BLS occupational wage release. The national mean for the same group is $56.16
2.9%
Share of metro employment sitting in computer and mathematical work, against 3.4% nationally. Chicago pays above the national average across all occupations combined. Not this group
17days
How long a Chicago AI search usually runs before the first names land in front of your panel
92%
Share of our Chicago-area AI placements still holding the seat a full year on
Every hiring manager who calls us about a Chicago AI role has already looked up a benchmark. Usually two or three. They arrive with a number. The number is real. It is also going to cost them a quarter.
Here is the one that matters most. Across the Chicago-Naperville-Elgin metro, computer and mathematical occupations average $52.08 an hour. Nationally that same group averages $56.16. Read that twice. The third-largest metro in the country, with roughly thirty Fortune 500 headquarters inside it, pays technical work under the national average while paying over the national average on everything else combined.
That gap is not a bargain. It is a warning label.
The reason is composition. Not stinginess. Computer and mathematical work is 2.9% of employment here against 3.4% nationally, and the bulk of it sits inside enormous non-technology employers, keeping claims systems and pharmacy systems and freight systems upright. Those are real jobs held by good people. None of them train models. Not one.
So the published benchmark describes a labor market you are not shopping in. Write your requisition to it and two things happen at once. The candidates you want never see the posting. The ones who do see it can’t do the work. Both halves fail at once.
The Benchmark
The Average Is Set by Jobs That Aren’t AI Jobs
Chicago runs on other people’s systems. Allstate in Northbrook, Zurich North America in Schaumburg, Health Care Service Corporation downtown, Abbott and AbbVie up in North Chicago, Baxter and Walgreens in Deerfield, McDonald’s in the West Loop, Grainger in Lake Forest, Discover in Riverwoods. Add Kraft Heinz, Mondelez, Conagra and US Foods and you have tens of thousands of technical seats holding enormous businesses together. Good work. Not AI work.
Those seats set the metro average. That is why it reads low. Nothing sinister about it.
Meanwhile the AI work in this city concentrates into a much smaller number of rooms, and those rooms pay on a different scale entirely. Precision medicine at Tempus. Model risk and pricing inside the carriers. Forecasting and routing across a freight and rail interchange that is the largest in North America. Research computing at Argonne National Laboratory in Lemont, where the Aurora exascale system runs 63,744 GPUs.
One metro, two wage populations, one published number that averages them together. Arithmetically honest. Operationally useless.
The Axis
Where Chicago AI Money Sits, and Where It Doesn’t
One scale. Base salary, running left to right. The blocks are the bands we see on live Chicago AI requisitions, which is recruiter observation rather than a published survey. The stretch between them is the part nobody puts in a benchmark, because there is nothing there to measure.
130–175K
Enterprise and carrier AI
Insurers, health systems, banks, CPG, logistics. Applied ML dropped into a business that already works. Largest employer group in the metro by a distance.
165–215K
Product AI at Chicago software firms
Health tech, legal tech, supply-chain visibility, marketing platforms. Smaller headcount. Faster loops. The top of this band is where you start losing people.
ContestedNo bids
Roughly 215K to 300K. Enterprise approval ceilings stop under it. Trading desks start over it. Nobody lives here. Almost nothing in this metro is priced in that stretch.
300K →
Trading and quant research
The Loop and the West Loop. Jump Trading, DRW, Optiver, IMC, Akuna, Belvedere, Wolverine, Chicago Trading Company, Citadel Securities. Total comp is not published, and the ceiling is not really a salary question. No ranges. No ladders.
First tick is the BLS metro benchmark for computer and mathematical work, annualized. The last one carries no number, because the firms at that end of the scale don’t publish one.
Brass is the instrument. The scale and its ticks. Orange is the contested band, the only stretch where both populations bid on the same person. If your offer tops out there, you are no longer competing with companies that look like yours.
The Contested Band
You Lose the Search at the Top of Your Own Range
Most Chicago AI searches don’t fail at the bottom of the band. They fail at the top, in the last two weeks, on the candidate everybody in the room had already agreed was the one. Late, and expensive.
Watch what happens to an infrastructure-shaped profile. Somebody who can serve a model under a latency budget, version it, roll it back and keep it alive at three in the morning is valuable to a hospital network and just as valuable to a market maker four miles away, and the two of them price that person against completely different internal ladders, one of which does not have a compensation committee attached to it.
We’ve watched that exact sequence twice this year in the West Loop. Both times, week seven.
The fix isn’t always more money, which is the part clients don’t expect. It’s usually deciding, before the posting goes up, whether you are hiring from the population that wants a defined scope and an interesting problem, or from the population optimizing a number. Different people. Different first email.
Roles
Six Chicago AI Titles and Who Else Is Bidding for Each
The band matters less than the competition for the profile. Far less. One of these six is wanted by both ends of the market at once, which is why it is the hardest of the six to close.
Machine Learning Engineer
Training, evaluation and the handoff to production. Here the model almost always lands inside an operating business with a compliance function and a change window.
LLM and Applied AI Engineer
Grounding a model in a corpus somebody already owns. Heaviest demand right now in legal tech, health tech and the carriers. Newest of the six.
ML Platform and MLOps Engineer
Serving, versioning, lineage, rollback. Wanted by a hospital network and a market maker on the same afternoon, at prices that don’t resemble each other.
The contested seatComputer Vision Engineer
Medical imaging along the North Shore, inspection and defect work in the manufacturing belt, yard and dock analytics across the freight corridor.
Quantitative Researcher, ML
Signal research on market data. Different interview, different comp structure, and a candidate pool that mostly never looks at a job board.
Applied Scientist
Forecasting, pricing, personalization, risk. Reports through a business line more often than through engineering, which changes who approves the offer.
Screening detail for each title lives on its own page. We go through interview structure and comp separately for ML engineers, for LLM and retrieval engineers, and for generative AI builders. All of it rolls up into our AI and ML practice, which sits on the technology staffing desk.
The Calendar
Chicago’s Best Supply Line Runs on an Academic Schedule
Two and a half hours south of the Loop, the University of Illinois turns out one of the largest computing cohorts in the country every spring. Northwestern, the University of Chicago, the Toyota Technological Institute and Illinois Tech add to it closer in.
That pipeline is enormous and it runs on a clock with nothing to do with your headcount approval. Firms that hire well out of it are talking to people in September for the following summer, which means a February requisition for a junior applied scientist is fishing in whatever is left after the coastal offers and the trading desks have already been through, and the people still available in February are usually available for a reason.
Mid-career hiring runs the opposite rhythm. Nobody moves in December. January and late summer are when the good ones look, and a search kicked off in the third week of December mostly buys you a slow month.
Plan campus-adjacent hires nine months out. Everything else moves around bonus season. Two calendars, one metro.
Coverage
Four Parts of Chicagoland, Four Different Offers
One search covers the whole metro. Every part of it prices differently. A candidate in Deerfield and a candidate in the West Loop are holding completely different alternatives while they read your offer.
01
The Loop, West Loop and Fulton Market
Exchanges, trading firms, product AI and the headquarters cluster. Highest ceiling in the metro and the shortest patience for a slow interview loop. Everyone here is contacted weekly.
02
The North Shore and Lake County
Abbott, AbbVie, Baxter, Astellas, Walgreens, Allstate. Medical imaging, drug discovery, actuarial modeling and claims. Regulated and deliberate. Far more willing to pay for somebody who has already shipped inside a validated environment.
03
The I-90 northwest corridor
Schaumburg, Hoffman Estates, Rosemont, Elk Grove Village. Insurance, telecom and distribution. Firm hybrid expectations. And a candidate pool that will not drive to the Loop three days a week for a lateral move.
04
The I-88 research belt
Naperville, Lisle, Warrenville, Batavia, Lemont. Argonne, Fermilab and the labs and manufacturers around them. Deep computational people. Long tenures. The part of the metro where a genuinely hard problem still outbids a raise.
Illinois also changed the rules underneath all four. The state’s amendment to the Human Rights Act, HB 3773, took effect January 1, 2026 and makes discriminatory AI in employment decisions a civil rights violation, calling out zip code used as a proxy. Illinois has also carried the strictest biometric privacy statute in the country since 2008.
Engagement
Three Ways to Buy the Same Desk
Same Chicago desk. Same candidate network. Three ways to paper it. What actually changes is who carries the risk while the seat is being filled.
Before the headcount clears
Contract and Contract-to-Hire
Our usual answer when the modeling work is funded and the requisition isn’t, or when watching somebody operate for a quarter is worth more to you than another interview loop.
Contract Staffing →A seat with a five-year shape
Direct Hire
Platform, governance and lead roles. The seats where whoever takes them answers for the system long after the current model is retired.
Direct Hire details →A bounded piece of work
Project and Statement of Work
An evaluation build, a migration off a vendor model, or a proof of concept that has to reach a decision by a date somebody already committed to.
Project Staffing →Common Questions
What does an AI engineer actually cost in Chicago?
Enterprise and carrier AI roles here generally land between $130K and $175K base, product AI at Chicago software firms runs $165K to $215K, and trading-side research sits far above both with no published number.
Those are the bands we see on live Chicago requisitions, not a survey. The BLS metro benchmark for computer and mathematical work annualizes near $108K, which shows you how far AI sits above the group it gets averaged into. Send the job description and the reporting line. What comes back is a single band for that exact role in that exact submarket, which is narrower than any range above and far more useful for a budget conversation.
Why is Chicago’s technical wage below the national average?
Composition. Computer and mathematical work is 2.9% of Chicago employment against 3.4% nationally, and most of it supports very large non-technology employers rather than building software products.
Chicago pays above the national average across all occupations combined, $34.42 an hour against $32.66. Chicago is not cheap. The technical average reads low because the denominator is packed with enterprise IT seats at insurers, distributors, health systems and manufacturers. Strip those out and the picture inverts, which is exactly why the published figure misleads people writing AI requisitions.
Do we have to compete with the trading firms?
For most roles, no. For platform and infrastructure profiles, yes, and it usually decides the search. That is the one band where a hospital network and a market maker want the same person.
Pretending otherwise is how good searches die in week seven. Two things reliably work. Scope the role so the interesting part is genuinely interesting, because plenty of strong Chicago engineers will take a real problem over a bigger number and will say so on the first call. Then move fast enough that your decision lands before the other conversation gets serious.
Did Illinois change anything about AI in hiring?
Yes. Since January 1, 2026 it is a civil rights violation in Illinois to use AI that discriminates in employment decisions, with zip code as a proxy called out specifically, and employers have to give notice that AI is being used.
Illinois has been ahead of most states on this for a while. The biometric statute has been in force since 2008 and carries the kind of exposure that makes a general counsel read your model documentation properly. Practically, it shows up in interviews. Candidates for governance-adjacent seats here get asked about it, and the ones who can answer without a lawyer in the room are worth more than they were two years ago.
Can we hire these roles remotely?
Most of them, yes. Plenty of Chicago AI engineers already work for companies headquartered somewhere else, and the metro exports remote technical talent more than it imports it.
That cuts both ways, and people forget the second half. Opening the role nationally means you are bidding against a coastal band instead of a Chicago one. Sometimes that is right. It is a decision worth making deliberately in week one rather than discovering from a counteroffer in week nine.
Where does Chicago AI talent actually come from?
Mostly from Illinois itself. The University of Illinois graduates one of the largest computing cohorts in the country, and Northwestern, the University of Chicago, the Toyota Technological Institute and Illinois Tech feed the metro directly.
Argonne National Laboratory in Lemont and Fermilab in Batavia add a deep computational population that almost never appears on a job board. Those people move for problems. Not postings. Reaching them is a different exercise from filling an enterprise ML seat and it takes longer, so budget the runway before you promise a date to your executive team.
How quickly do we see candidates?
Seventeen days is the typical Chicago run from intake call to a first qualified shortlist, and a year later 92% of those hires are still in the seat. Enterprise ML and applied science roles often beat it.
Platform and quant research are slower. Much slower. Those populations are tiny and effectively all of them are employed, so the work is persuasion rather than sourcing. We tell you which of the two you have during intake, not five weeks in. Name the end of the market you intend to hire from when you send the role over, and the date we give you back will be one you can defend to your leadership.
Pick your end of the scale before you write the requisition. We’ll tell you who is reachable there.
Send the role and the band you have approved. You get back an honest read on who is genuinely available at that number in Chicago, who else is bidding for them, and what it takes to make them answer.
Get a Chicago AI Shortlist →