Software Engineer - Machine Learning at FocusKPI Inc.
Mountain View, CA
$95–$110/hrfrom the description
Jun 30, 2026
Mountain View, CA
Jul 21, 2026
What this job asks for AI summary
This role centers on building and maintaining safety infrastructure for agentic AI systems — specifically, models that detect prompt-injection attacks and unsafe content across mobile, cloud, and XR/AR environments. Day-to-day work involves training and deploying classifier and guardrail models, applying post-training techniques such as RLHF and DPO, curating adversarial datasets, and integrating safety pipelines with device and platform teams. It suits an ML engineer with hands-on production experience in LLM post-training and safety or content-moderation systems.
Mid level · 3+ years · San Jose-Sunnyvale-Santa Clara, CA · Contract
“or” means any one of them counts — you don't need all of them.
We read this from the posting text with AI. Skim the description below before ruling yourself out.
How this req sits in the market our data
Roughly 20 people in the San Jose-Sunnyvale-Santa Clara, CA area plausibly meet what this posting asks for (data scientists). range 6–35
Applicant volume Moderate — A normal amount of company. The rare requirements below are what will separate a shortlisted application from the rest.
Rare in this occupation — lead with these, and say what you built with them.
Most people in this occupation already list these. Still required — just not what gets you shortlisted.
What the occupation pays Median $189,150 (middle half $152,859–$224,286). This posting is about at that midpoint.
Estimated from BLS employment for this occupation and area, per-skill prevalence across our listing corpus, and published wage benchmarks — as of Jul 28, 2026. It is a model, not a headcount.
Why we read it this way (10)
The role title is 'Software Engineer - Machine Learning' but the primary day-to-day work is training, post-training, and deploying ML/AI safety models (classifiers, guardrails, RLHF/DPO optimization) — closer to Data Scientists (15-2051) than general Software Developers (15-1252), though the boundary is genuinely ambiguous; alt SOC 15-1252 is noted.
The degree requirement is nuanced: the JD lists M.S. or Ph.D. as the primary education path but explicitly allows a B.S. with 'equivalent industry experience' — so no hard degree gate applies.
The 3+ years of ML engineering experience is stated post-master's degree; the JD also accepts a Ph.D. with 1 year post-graduation, making the effective minimum experience variable. 3 years is used as the overall years minimum as the most commonly applicable floor.
DeepSpeed, FSDP, and Accelerate are listed together as distributed training frameworks under the required qualifications section; DeepSpeed is used as the primary skill with the others as alternatives since all three serve the same function.
RLHF, DPO, and RLAIF are listed as interchangeable post-training techniques under a single requirement; RLHF is used as the primary with DPO and RLAIF as alternatives.
On-device/edge deployment tools (ExecuTorch, Core ML, TFLite, MLC-LLM) and model compression techniques appear under Preferred Qualifications.
Red-teaming, adversarial data generation, and automated attack pipelines appear under Preferred Qualifications.
This is a 12-month W2 contract (C2C explicitly excluded); classified as Contract accordingly.
Mountain View, CA falls within the San Jose-Sunnyvale-Santa Clara, CA CBSA (41940).
Posting is for a contract engagement — the market benchmarks below price full-time roles, so read the comp comparison with that in mind.
Read the full posting
The employer publishes the full description on their own site — read it there ↗. Or sign in to read it here — it's free, and it also lets you track this application.