Applied AI Engineer

DeepRec.ai; España; Remote; EUR 90,000 - 130,000; Full time

Job summary

DeepRec.ai is hiring an Applied AI Engineer to join a small founding team building an AI native productivity platform. The role focuses on creating reliable ML-powered features across email-oriented workflows, with an emphasis on context maintenance, decision support, and user-in-the-loop approvals.

You will design prompts, tools, memory and agent workflows; train and deploy models using PyTorch or JAX, LLaMA/Qwen, and vector databases, with a remote-first setup across Europe.

Qualifications

• Demonstrated depth in building production ML systems.
• Ability to deploy ML models end-to-end.
• Experience with prompts, tools, memory and agent workflows.
• Ability to optimize ML systems for latency, cost and reliability.

Responsibilities

• Build AI features from model through to user experience.
• Train, fine tune, evaluate and deploy machine learning models.
• Design prompts, tools, memory and agent workflows.
• Improve accuracy, latency, cost and production reliability.
• Debug issues across models, orchestration, infrastructure and product.

Skills

ML development

Model deployment

Prompt design

Memory & agent workflows

Debug ML systems

Tools

PyTorch

JAX

LLM APIs

LLaMA

Qwen

vLLM

Vector databases

Job description

I am hiring an Applied AI Engineer to join a small founding team building a new AI native productivity platform.

The first application is rethinking how people manage email.

Rather than asking users to spend hours reading, sorting and replying to messages, the platform is designed to:

• Prioritise and classify emails across multiple inboxes


• Draft replies and prepare summaries


• Identify tasks and calendar actions


• Complete multi step workflows using external tools


• Remember preferences and previous context


• Ask for approval before sending messages or taking actions

The aim is to reduce the average time spent managing email from around four hours per day to approximately thirty minutes.

This is a proactive AI product built for people who do not want to learn complex prompting. The system should understand context, prepare work and present clear decisions for approval.

The technical challenge is reliability.

The platform needs to maintain context across long running workflows, produce structured and predictable outputs, recover from errors and complete real tasks without removing user control.

This is not an API wrapper role.

The team needs someone with genuine machine learning depth who can take model capabilities from research and experimentation into reliable production systems.

You will:

• Build AI features from model through to user experience


• Train, fine tune, evaluate and deploy machine learning models


• Design prompts, tools, memory and agent workflows


• Improve accuracy, latency, cost and production reliability


• Debug issues across models, orchestration, infrastructure and product

The core stack includes Python, PyTorch or JAX, LLM APIs, open models such as LLaMA and Qwen, vLLM and vector databases.

The opportunity is remote first, with candidates being considered across the United Kingdom and Europe. The package includes competitive cash compensation and equity, structured around the individual.

The process begins with a monitored Python and machine learning assessment, followed by technical and founder discussions.

If you have built real ML systems in production and want significant ownership over a product moving from prototype toward launch,