Back to jobs

Applied LLM Engineer

Pay
Pay not listed
Work setting
Remote — work setting source
WHO THIS IS FOR You break things on purpose and explain exactly why they broke. You can read a bad answer and say in one sentence what is wrong with it, and read a trace and say in one sentence why it was slow. You would rather run fifty tests than win one argument about which prompt is better. You know what a turn costs, and you write well enough that people notice. Location: Remote, global Work format: Remote, full time WHAT YOU WILL DO Prompts and evaluation (45%)
Read the full posting
Employment
Full-time — employment source
Employment type in the employer’s posting: Full-Time
From the employer’s posting

What you’ll work on

Full posting

House of YO is a connectivity and distribution business with compute layered on top.

On V3 of the YO platform, YOlanda is the interface.

Apply at Yomobile

Posting record

Report an error
Status in our records
Live
First seen by us
Oct 22, 202630 days ago
Recorded sightings
29
Last seen by us
Nov 21, 2026

These dates show when we found the posting. Check the employer’s website to confirm it is still accepting applications.

How this job compares

Time since we first saw it

Day 30

Up longer than most. Similar jobs usually come down between day 1 and day 15.

From 1,368 similar postings that came down. Coming down is not always a hire.

Yomobile is hiring

2 open jobs

Open jobs by the week we first saw them, last 12 weeks.

Hiring record

See how this role fits your experience

Add your resume and we’ll mark which of these tools you already use, and point out anything in the posting that could rule you out.

  • Python
Where the posting names them
- Two or more years building with large language models in production, at real traffic. Daily use of a chat window does not count. - Python you can ship. Scripts, API clients, data handling, tests. Git, JSON, regex, markdown. - LangChain and LangGraph. You have shipped with both, and you can read a graph you did not write and change it without breaking it.

Your next step

Complete your application on yomobile.bamboohr.com. The employer’s form will show what is required.

Applying through BambooHR

  • No account needed. The form takes images too, but an image of a resume has no text for anyone to search: upload a PDF or Word file.

Already applied? Track this application

Job description

View original posting ↗

ABOUT THE ROLE 

House of YO is a connectivity and distribution business with compute layered on top. YOlanda is our AI concierge.

On V3 of the YO platform, YOlanda is the interface. Users talk to her instead of tapping through menus. She answers questions about plans, data, top-ups and YOYO$ balances, and she takes the action on the user’s account. When she is wrong, the user is stuck and the product has failed.

Her knowledge is the YO ecosystem and nothing else. When a question falls outside it she hands off to YOnC, our compute cooperative, which routes it to a specialist model. Knowing where that line sits, and holding it, is central to the job.

She works to one golden rule: within five responses she has put a product or service in front of the user. Not five responses to be helpful. Five to arrive somewhere real.

This role makes her right, and keeps her fast, affordable and accountable in production. You write the prompts that govern how she thinks, build the evaluations that prove she works, run the caching, routing and controls that let her do it at scale at a cost per turn we can defend, and ship the code behind all of it.

YOlanda runs on our own LangChain and LangGraph build. You will open pull requests against that graph alongside the two engineers who own it.

You will not manage anyone. You will be in the model and in production every day.

WHO THIS IS FOR
You break things on purpose and explain exactly why they broke. You can read a bad answer and say in one sentence what is wrong with it, and read a trace and say in one sentence why it was slow. You would rather run fifty tests than win one argument about which prompt is better. You know what a turn costs, and you write well enough that people notice.

Location: Remote, global
Work format:
Remote, full time

WHAT YOU WILL DO
Prompts and evaluation (45%)

  • Own YOlanda’s system prompt. Persona, tone, safety rules, what she refuses, when she hands off.
  • Write and maintain the prompts behind every user action on V3: plan changes, top-ups, balance queries, offers, troubleshooting, account support.
  • Hold the golden rule. Design the conversation so she reaches a product or service within five responses, without becoming pushy. Every clarifying question costs a turn she cannot get back.
  • Own the boundary. Decide what is inside YOlanda’s frame of reference and what goes to YOnC, and make the handoff seamless: same conversation, same voice, the answer simply arrives.
  • Manage context. Decide what YOlanda knows about the user, the account and the plan catalogue at each step, and what she must not carry over. Break long tasks into chains and tool calls.
  • Build the evaluation set for YOlanda. Every task she performs gets test cases, a pass condition and a score you can track over time, in both languages.
  • Measure the boundary, fact check plan prices, data allowances and YOYO$ maths, and red team the top-up and rewards flows. A confident wrong number costs a user money.

Production, caching and optimisation (35%)

  • Trace every call. Prompt version, model, tokens, cache hits, latency and cost on every turn. Set the SLOs for latency and availability, and build the alerts.
  • Own the gateway behaviour YOlanda depends on: timeouts, retries, and fallbacks across providers and models when one degrades.
  • Write prompts for the cache. Stable parts first, in a fixed order: system prompt, tool schemas, plan catalogue, then the user’s turn. A prompt that cannot be cached costs money on every turn.
  • Build semantic, retrieval and tool-result caching, scoped and invalidated correctly. A cached price from last week is a hallucination with a timestamp. A cached answer about someone else’s balance is a breach.
  • Own cost and latency per turn. Token budgets, context trimming, output length, and batch APIs for offline work.
  • Route by task. Benchmark frontier and open models on accuracy, latency and cost per turn, choose the model per task, and bring the numbers. Work with the engineers running open models on YOnC compute on what they serve and how.

Governance (20%)

  • Keep prompts, models and configuration in git with the rest of the code. Versioned, reviewed, tested, owned, and rolled back in one step when needed.
  • Build the release gate in CI: evaluation score, cost and latency budget. Then a canary, a staged rollout and a kill switch.
  • Audit every action that touches an account or money, from the user’s message to the tool call. Redact personal data before it reaches logs or a third-party provider.
  • Put the guardrails in the runtime: prompt injection defences, tool permissions, spend limits on top-up and rewards, and the point where a human approves.
  • Set cost budgets per feature and per model, with alerts. Know the rules that apply when an AI handles money and an operator account, and turn them into controls.

WHAT YOU NEED

  • Two or more years building with large language models in production, at real traffic. Daily use of a chat window does not count.
  • Python you can ship. Scripts, API clients, data handling, tests. Git, JSON, regex, markdown.
  • LangChain and LangGraph. You have shipped with both, and you can read a graph you did not write and change it without breaking it.
  • Prompting and evaluation. System prompts you wrote that shipped, and the evaluation harness that proved them. Code or work we can read. You change one variable at a time and record the result.
  • Production operations. Tracing, SLOs, fallbacks, and caching you have shipped with invalidation that was right. A before and after on cost and latency, with numbers on it.
  • Governance in practice. Versioning, release gates, audit logging, personal data handling. Controls in code, not policy in a document.
  • Retrieval in production. Chunking, embeddings, reranking, grounding an answer in a source, and knowing when retrieval is the wrong tool for the question.
  • Exceptional written English. Syntax, register, nuance. You can write in a voice that is not your own and hold it.
  • Care for the user at the other end. Many of our users are on prepaid plans and count every dollar. YOlanda gets it wrong at their expense, not ours.


WHAT WE OFFER

  • Top-notch products disrupting the world of the traditional media-services industry
  • A brilliant, highly collaborative team with a shared drive to achieve goals
  • Flexible working hours
  • Competitive salary
  • Real opportunity for personal and professional growth

Demand for AI Engineer roles

New AI Engineer postings at the employers we follow are up 340% over the last 6 weeks compared with the 6 before.

Weekly new AI Engineer postings at the same employers, from the week of August 17 to the week of November 9. The fewest in a week was 78, the most 803, and the latest week had 438.
August 17Last week

Each bar is one week of new postings. The darkest bars are the last 6 weeks.

Counted from the hiring pages we’ve followed for at least five weeks, so employers we add don’t look like growth. Updated weekly.

About the role

We file this job under AI Engineer, by the work the posting describes.

AI engineers build products on top of AI models that already exist, most often large language models: they connect a model to a company's data and tools and ship the assistants, agents and search features people use.

Postings ask for strong Python or TypeScript, experience with model APIs, retrieval and agent frameworks, and ways to test whether the model's answers are good enough.

An ML engineer trains and serves models; an AI engineer usually starts from a model someone else trained and builds the product around it, though titles often mix the two.

Source details

View original posting ↗

Source notes

Source excerpts

Selected passages from the saved posting. Check the full description for conditions and exceptions.

Pay

No pay amount identified in the saved description.

Location & working pattern

Location not supplied.

WHO THIS IS FOR You break things on purpose and explain exactly why they broke. You can read a bad answer and say in one sentence what is wrong with it, and read a trace and say in one sentence why it was slow. You would rather run fifty tests than win one argument about which prompt is better. You know what a turn costs, and you write well enough that people notice. Location: Remote, global Work format: Remote, full time WHAT YOU WILL DO Prompts and evaluation (45%)
Work authorization

No clear work-authorization passage found. Eligibility is unconfirmed.

Find answers in the posting

AI
How answers work

AI selects complete passages from this posting. Check them for conditions and exceptions.

Uses this posting and your question. No profile needed.