ML Engineer at AdTech and Programmatic Advertising Companies: H-1B Sponsorship Reality

AdTech companies like The Trade Desk and Criteo sponsor ML engineers on H-1B, but the process has specific quirks you need to know before you apply.

You applied to a machine learning role at a programmatic advertising platform. The recruiter sounds excited. The tech stack is real — bidding algorithms, real-time feature pipelines, CTR prediction models. But somewhere between the phone screen and the onsite, a familiar question surfaces in your mind: does this company actually sponsor H-1B? And if they do, what does that path look like for someone on F-1 OPT?

AdTech is a legitimately strong sector for international ML engineers. The industry runs on data, and the people who build models that make that data commercially valuable are in genuine demand. But adtech spans public companies with dedicated immigration teams to seed-stage startups that have never filed an I-129. Knowing which side of that line a company sits on — and what the sponsorship path looks like — is the intelligence that shapes whether you get a visa or get stuck.

What AdTech ML Engineering actually looks like

ML engineers at adtech and programmatic advertising firms typically work on:

• Real-time bidding (RTB) models — predicting CTR and conversion rate within milliseconds per impression
• Audience segmentation and lookalike modeling — clustering and classification at scale across first and third-party data
• Feature engineering pipelines — data infrastructure feeding model training from impression logs and bid streams
• Optimization and ranking — budget pacing, multi-armed bandit experimentation, auction mechanics
• Privacy-preserving ML — differential privacy and clean-room analytics as third-party cookies disappear

These roles require applied statistics, machine learning, and distributed systems — they sit squarely in the specialty-occupation bucket for USCIS. A petition framing the role as "data analyst" rather than "ML engineer" is where trouble starts.

For broader context, see our general ML engineer H-1B guide and the data science sponsorship overview for 2026.

Which adtech companies actually sponsor

The adtech ecosystem ranges widely in size and immigration infrastructure. Here is a practical breakdown:

The Trade Desk is frequently cited as one of the more international-friendly adtech employers for technical roles. They appear consistently in USCIS H-1B data year over year with approvals for ML, data engineering, and software engineering titles. That said, they are a cap-subject employer, so you still go through the lottery.

To verify any specific company before you apply, use the USCIS H-1B Employer Data Hub and filter by company name. Look for approval volume, denial rate, and whether they file for ML-adjacent job titles.

Your status roadmap — from OPT to H-1B

Phase 1: OPT (12 months)

The moment you graduate and activate your F-1 OPT, you have 12 months to work for a qualifying employer. This period starts your first real opportunity to land an adtech ML role and demonstrate value before the H-1B lottery.

Key rules during standard OPT:

• Employment must be directly related to your field of study
• You cannot have more than 90 cumulative days of unemployment during OPT
• Your employer does not need to sponsor anything during this phase; your authorization comes from your OPT EAD card

Phase 2: STEM OPT extension (24 months)

If your degree was in a STEM-designated field — which includes computer science, data science, electrical engineering, statistics, and applied mathematics among many others — you can apply for a 24-month extension. This brings your total authorized work period to 36 months.

During STEM OPT your employer must be E-Verify enrolled, sign an I-983 training plan with your DSO, and report employment changes within five business days. The 90-day unemployment limit carries across the full OPT + STEM OPT period — adtech hiring can freeze during ad-market downturns, so track your days carefully.

For more detail, see OPT vs STEM OPT vs CPT.

Phase 3: H-1B lottery and cap-gap

The H-1B lottery registration window opens in early March each year for an April 1 start date. USCIS first selects from a master pool (all registrations), then from a sub-pool of master's-degree holders from US universities. If your field of study is STEM and you have a US master's, you get two draws — statistically better odds than a single draw.

If selected, your employer files a full I-129 petition. If you are on STEM OPT when the October 1 H-1B start date arrives, the cap-gap rule extends your F-1 status and work authorization through September 30, protecting you during the transition period. Under the H-1B Modernization Rule effective January 2025, this cap-gap protection was extended through April 1 of the relevant fiscal year, giving additional runway.

If you do not get selected in the lottery, you need a backup plan. See H-1B backup plans after the lottery for the full menu of options.

Prevailing wage and what it means for ML roles in adtech

The Department of Labor sets prevailing wages for H-1B positions through a four-tier system (Level I through Level IV). Your employer's LCA (Labor Condition Application) must state a wage at or above the DOL prevailing wage for the job title and location.

For ML engineer roles in major adtech markets (New York, San Francisco, Los Angeles, Seattle), prevailing wages tend to be set at Level II or Level III, reflecting the experience typically required. This is generally not a problem at well-funded adtech companies, whose market salaries often exceed prevailing wage levels comfortably. Where it becomes an issue is at cash-constrained startups that want to offer below-market salaries — USCIS will scrutinize those petitions.

One thing to verify: the job title on your LCA should reflect your actual duties. A petition filed as "Software Developer" for a role that is substantively "Machine Learning Engineer" can invite an RFE asking for evidence that the position qualifies as a specialty occupation in that specific classification.

Green card strategy in adtech

Starting your green card process early matters, especially if you are from a high-backlog country.

For candidates from India and China

EB-2 and EB-3 PERM backlogs for Indian nationals are severe. This is the single biggest career-planning risk for Indian and Chinese ML engineers at adtech companies.

The actionable alternatives:

1. EB-1A (Extraordinary Ability) — No PERM, no employer sponsorship required. Qualifying criteria to build toward: peer-reviewed publications in ML/ad systems, invited talks at conferences (NeurIPS, RecSys, AdKDD), significant open-source adoption, and media coverage. Years of intentional effort, but the most reliable path around the backlog.

2. EB-2 NIW (National Interest Waiver) — Self-petition without PERM. ML engineers working on privacy-preserving ML or fraud detection have made successful NIW arguments. See our EB-2 NIW self-petition guide for the framework.

For candidates from other countries

Standard EB-2/EB-3 PERM typically takes two to four years for non-India, non-China nationals. Start as early as your employer will allow — the priority date is set when PERM is filed, not when the green card is issued.

The NLP angle — programmatic advertising ML and language models

A growing subset of adtech ML work involves natural language processing — classifying ad content, contextual targeting without cookies, brand safety classification, and ad copy generation. If your background is in NLP or LLMs, this expands your target role pool. The role description also benefits from the added specificity of NLP skills, strengthening the specialty-occupation argument. Our NLP engineer H-1B guide covers the broader picture.

How to evaluate a specific adtech company

Before you invest heavily in an interview loop, run this checklist:

1. Check USCIS H-1B Employer Data Hub — search the company name, look at approvals vs. denials, and confirm job titles match ML engineering.
2. Check E-Verify enrollment — required for STEM OPT; if they are not enrolled, the role does not work for you.
3. Ask the recruiter directly — "Do you have immigration counsel on retainer, and have you sponsored ML engineer roles previously?" Hesitation here is a yellow flag.
4. Look at LCA disclosure data — myvisajobs.com aggregates wage levels and job titles from DOL filings for every employer.
5. Research financial stability — H-1B requires proof of ability to pay the prevailing wage; a cash-strapped startup mid-petition is a real problem.
6. Review tenure on LinkedIn — adtech layoffs track the ad market. If most ML engineers left within two years, understand why before you sign.

For startup-specific risk, see can this startup sponsor H-1B.

Common mistakes

Starting the interview process without checking sponsorship history. It takes three minutes on the USCIS data hub. Spending three weeks in an interview loop at a company that has never filed an H-1B is time you cannot recover.

Letting the "data analyst" title slide. Adtech companies sometimes bucket ML work under analyst titles to manage headcount. A job titled "Data Analyst" on your H-1B petition faces much higher specialty-occupation scrutiny than "Machine Learning Engineer." Push for the accurate title before your offer letter is finalized — changing it afterward is harder.

Ignoring the 90-day OPT unemployment limit during job search. AdTech hiring can freeze during ad market contractions (Q1 typically slows). If you are searching for your first role or between roles during OPT, count your unemployment days. Beating the OPT 90-day clock walks through strategies to stay compliant.

Not starting green card conversations with your employer early. Many adtech companies will start PERM once you have a year of tenure if you ask. The candidates who end up deepest in the backlog are often the ones who waited until year three or four to start. The I-140 priority date is set by when PERM is filed, not when the green card interview happens — starting earlier buys you years of queue position.

Assuming The Trade Desk or Criteo means automatic approval. Established employers significantly reduce petition risk, but USCIS still reviews each petition on its merits. A poorly packaged petition with a vague job description, a wage level that does not match actual duties, or missing supporting evidence can draw an RFE even from strong employers.

Overlooking cap-exempt options as a bridge. If you do not get selected in the H-1B lottery, a university or nonprofit research role — even a short-term appointment — can serve as a bridge while you re-enter the lottery next year. Cap-exempt employers include universities, nonprofit research organizations, and government research facilities. See cap-exempt H-1B employers for the full list.

Skipping premium processing. At $2,965, premium processing is real money, but it guarantees adjudicative action within 15 business days. For H-1B transfers between adtech employers, or for cap-year filings close to your OPT expiration, the certainty premium processing provides is worth it.

Frequently asked questions

Do adtech companies like The Trade Desk sponsor H-1B visas for ML engineers?

Yes. The Trade Desk, Criteo, LiveRamp, and others have consistent multi-year H-1B filing histories. They are cap-subject employers, so ML engineers on OPT must enter the lottery. Verify any company's record on the USCIS H-1B Employer Data Hub or myvisajobs.com before investing in an interview loop.

How does OPT and STEM OPT work for ML engineers at adtech companies?

Standard F-1 OPT gives you 12 months of authorized work. A STEM degree (CS, data science, electrical engineering) qualifies you for a 24-month extension, totaling 36 months. Your employer must sign an I-983 training plan during STEM OPT. You cannot accumulate more than 90 days of unemployment across the combined period without risking your status.

What does the H-1B specialty-occupation requirement mean for an ML engineer role at an adtech company?

USCIS requires that the role normally demand at least a bachelor's in a specific field. ML engineering at adtech companies — requiring applied math, statistics, and CS — generally qualifies. The risk is a vague petition that describes the work as "data analysis" or "marketing analytics" rather than machine learning engineering, which invites an RFE.

What are realistic green card timelines for ML engineers at adtech companies from India or China?

EB-2 and EB-3 PERM backlogs for Indian nationals are severe — potentially decades at current priority dates. EB-1A (extraordinary ability, no PERM) and EB-2 NIW (national interest waiver, self-petition) are the actionable alternatives. Candidates from countries other than India and China typically wait two to four years under EB-2/EB-3.

Can a startup adtech company sponsor my H-1B visa?

Yes, but USCIS scrutinizes small companies on ability to pay prevailing wage and genuine employer-employee relationship. Check the company's H-1B history, confirm they have experienced immigration counsel, and use our startup H-1B checklist before signing an offer.

AdTech is a real and underappreciated sector for international ML engineers — the data volumes are large, the modeling problems are genuinely interesting, and the established players have the institutional infrastructure to run H-1B petitions properly. The risks are real too: cap-subject lottery exposure, specialty-occupation framing on the petition, green card backlog for high-demand countries, and startup instability in a sector that tracks the ad market. None of those risks are unique to adtech, and none of them are unmanageable with the right preparation.

If you want help thinking through your specific situation — your degree, your current status, your target employers, and your green card timeline — F1Jobs works with international ML engineers navigating exactly this path every day.