Senior AI Scientist – Model Training & Architecture (m/w/d)

We're building the operating system for the next generation of computing — one where AI agents replace apps and your technology finally works for you instead of the other way around.

We're a stealth-mode startup with a world-class founding team with deep roots in consumer AI, extended reality, and wearable technology — including founders of some of the most recognizable hardware and software platforms of the last decade. We're backed by strategic partnerships with leading silicon and manufacturing companies, and we're hiring our first AI engineer to build the intelligence layer at the core of the platform.

This is a rare opportunity to architect the agent infrastructure of a platform that doesn't exist yet — at the layer where always-on contextual AI meets a wearable form factor for the first time.

Additional product details shared under NDA.

What We Offer

• Salary: competitive depending on experience
• Meaningful early-stage equity
• Full medical, dental, and vision coverage
• Fully remote with occasional in-person time in Silicon Valley or Paris for key milestones

Awear is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.

Key Responsibilities
The backend engineers build the infrastructure. The mobile engineers build the user facing surfaces. You build what runs between them — the agents themselves.

As our first AI Engineer you will own the design and implementation of our agent layer — the pipelines, reasoning chains, memory retrieval systems, tool integrations, and orchestration logic that turn raw LLM capability into a platform that genuinely replaces the app paradigm. You will work directly with the CEO and across the full engineering team to make sure the agent experience is as technically rigorous as it is experientially compelling.

This is a hands-on engineering role. You will write production code, own the agentic runtime architecture, and be directly accountable for the quality of every agent interaction on the platform. You will also be a key voice in decisions about which models to use, how to route between them, and how to structure the memory and context systems that make our platform smarter over time.

We actively use AI development tools across our engineering team — Cursor, Claude, Copilot — and expect engineers who use them seriously as a core part of their workflow.

What You'll Build

• The platform agent runtime — the core orchestration layer that manages agent sessions, chains reasoning steps, routes to tools, and executes actions on behalf of users
• Multi-provider LLM integration and routing — selecting and switching between regional and task-specific language models dynamically, with latency, cost, and capability all factored into routing decisions
• RAG architecture and memory retrieval — the systems that give agents access to the user's persistent, encrypted context layer and make responses smarter and more relevant over time
• Tool and skill integration pipelines — the infrastructure that connects agents to external APIs, device capabilities, and first-party platform features
• Agent evaluation and observability — the frameworks that measure agent quality, surface failures, and give the team visibility into how agents are actually performing in production
• On-device inference optimization — working with the mobile and firmware teams to identify which parts of the agent pipeline can run locally on device, reducing latency and cloud dependency as the platform evolves toward wearable hardware
• Prompt architecture and system design — the structured prompting frameworks, system instructions, and context management patterns that govern agent behavior consistently across the platform

Ideal Experience

• 4+ years of software engineering experience with at least 2 years working directly on LLM-based systems in production
• Deep hands-on experience with LLM integration — not just API calls but genuine understanding of how to build reliable, scalable, low-latency AI pipelines
• Strong experience with RAG architectures — vector databases, embedding pipelines, retrieval optimization, context window management
• Familiarity with agent frameworks and orchestration patterns — LangChain, LlamaIndex, AutoGen, or similar, with a clear point of view on their strengths and limitations
• Solid Python engineering skills — you are writing production code, not research notebooks
• Experience evaluating LLM outputs at scale — building evals, measuring quality, detecting regressions
• Genuine understanding of the tradeoffs between different LLMs — capability, latency, cost, privacy implications — and experience making routing decisions in production systems
• Active user of AI-assisted development tools with a genuine point of view on how to use them well
• Strong written English and proven ability to work effectively in a remote and distributed team
• Strong plus: experience with on-device or edge inference — Core ML, ONNX, TensorFlow Lite, or similar; multi-agent system design; always-on or streaming inference architectures; privacy-preserving AI systems; wearable or mobile AI platform experience

TL;DR

We're hiring a Value Engineer to guide public sector customers from first conversation to long-term success. You'll help Sales to qualify and win the right deals, design solutions and services engagements that work in the real world. The objective is to ensure customers reach value quickly, own the coordination of all resources required for value realization, and proactively create opportunities to grow the relationship.

This role is a prescriptive, opinionated authority on which AI use cases make both business and technical sense. You'll develop a structured account thesis from day one and tie every decision back to measurable outcomes.

Why deepset

At deepset, we're on a mission to make custom AI solutions accessible to every organization. With Haystack, thousands of developers build advanced LLM applications every day, while our enterprise-ready AI Platform helps companies turn large language models into business value. We're remote-first, flexible, and built on a culture of trust and ownership. You'll collaborate with top-tier tech talent, tackle meaningful challenges, and help transform complex AI into solutions that are simple, powerful, and ready for the real world.

What you will do

• Value Discovery: Understand the Customer's Needs
• Lead structured, multi-persona discovery to uncover goals, challenges, KPIs, and decision-making dynamics across all relevant stakeholders; from end users and Line of Business owners to technical architects and executive sponsors.
• Identify the full landscape of stakeholder needs; distinguishing between technical requirements, business outcomes, and organizational constraints; being able to adjust pitch, messaging, and solution positioning autonomously per persona.
• Act as a sparring partner with Enterprise Sales on deal quality, qualification rigor, and expansion potential, with a shared accountability for revenue closure.
• Solution Design: Creating the Right Solution
• Translate customer needs into clear technical and business solutions, and define the full offering required for the customer to succeed; including scoped services engagements covering effort, team and project structure (including partner involvement if required).
• Build demos and POCs that prove value against defined business metrics, not just technical feasibility.
• Own AI risk management across project feasibility, data evaluation/readiness and solution adoption.
• Value Delivery: Own the Path to Customer Value
• Act as the orchestration layer above the project lifecycle and collaborate with our Technical Project Managers and Solution Engineers to:

• Define the rollout structure, success conditions, and resource accountability; ensuring the delivery team executes against them.
• Drive the plan to "first value" and own the broader rollout plan beyond it, including phased milestones, go-live readiness criteria, and post-launch stabilization.
• Coordinate customer and internal teams (FDEs, CS, Partners) so delivery stays aligned to the value commitments made in pre-sales.
• Monitor adoption KPIs and strategic success metrics defined at the outset, and surface deviations early.

• Proactively identify blockers across the full success ecosystem: missing integrations, end-user experience gaps, training deficits, organizational change barriers, and partner capability needs; and take ownership of resolving or escalating them.
• Ensure documentation, runbooks, and training plans are in place so customers can operate independently and achieve sustained adoption.
• Value Realisation: Measure Impact & Ensure Success
• Track usage, performance, and business outcomes against the success criteria defined in the account thesis.
• Run structured check-ins and QBRs that actively articulate the value delivered and reinforce deepset's differentiation; not just status updates.
• Continuously narrate value in a way that maintains executive alignment and competitive positioning.
• Resolve escalations and ensure systemic issues are addressed, not just patched.
• Value Expansion: Grow the Account
• Build and maintain a use case pipeline for each account: identify the next two to three use cases beyond the current scope, prioritize by feasibility and business impact, and develop a plan to advance them.
• Drive urgency with the customer through executive engagement, milestone anchoring, and proactive surfacing of opportunity cost of inaction.
• Run structured product update cadences to keep customers informed of roadmap progress, new capabilities, and relevant beta opportunities; positioning deepset's evolution as a competitive advantage.
• Use peer benchmarking and cross-customer case studies to inspire expansion, validate investment, and reinforce the customer's confidence in their AI strategy.
• Partner with Sales to build expansion value cases and support forecasting, with a clear handoff: the VE creates and qualifies the opportunity; Sales owns the close.
• Represent the Customer Internally
• Bring structured, signal-rich feedback to Product and Engineering that reflects real deployment experience, not just feature requests.
• Balance breadth of customer signal with prioritization judgment, distinguishing strategic product gaps from one-off edge cases.

What you bring to the team

• Software engineering background with significant experience in Sales Engineering, Solution Consulting, or Technical Implementation.
• Autonomous commercial capability and the ability to craft and deliver tailored value messaging, position deepset's offering against alternatives, and operate as a revenue-contributing partner to Sales without requiring hand-holding.
• Competitive positioning instinct and the ability to articulate deepset's differentiation clearly, handle objections, and use customer evidence and peer benchmarks as commercial tools.
• Practical Python skills for scripting, prototyping, and troubleshooting.
• Experience with AI/ML workflows, such as fine-tuning, data preparation, model evaluation, and RAG pipeline design.
• Strong understanding of modern architectures: APIs, integrations, IAM/security; bonus for Kubernetes, Terraform, SSO, and VPC.
• Ability to get hands-on with data, SQL, and light integrations.
• Strong orchestration and project leadership skills: ability to coordinate cross-functional teams (TPMs, FDEs, Partners) effectively, while maintaining ownership of value outcomes.
• Executive-level communication, with strong structured storytelling for C-suite and LoB audiences, not just technical audiences.
• Commercial awareness: services scoping, ROI framing, and a focus on measurable outcomes.

Nice to have

• Demonstrates strong systems thinking paired with tactical execution, with sound judgment to choose the right approach at the right time.
• Experience engaging with or selling into public sector organizations, with familiarity with procurement cycles and compliance constraints.
• Prior experience working with or alongside a partner ecosystem to extend delivery capability.

Benefits

• Remote-first setup with flexible hours & tech of your choice
• 30 days vacation + extra days for family sick leave
• Competitive salary & stock options for every team member
• Monthly sports & mental health support allowance with Oliva
• Annual learning & development budget
• Monthly team socials & in-person meetups
• Dog-friendly Berlin HQ

About Us
Founded in 2018, deepset builds open and enterprise-grade tools that help teams build AI with purpose. From Haystack, our open-source framework, to the Haystack Enterprise Platform, we give developers and organizations the building blocks to solve complex, high impact challenges with AI with full control, transparency, and sovereignty. Backed by GV and Balderton, we’re growing the world’s production AI community and customer base solving challenges too critical to get wrong.

‍
Visit us to learn more:
deepset Website | Haystack Website | GitHub | Linkedin | X deepset (Twitter) | X haystack (Twitter)

🎤 Why voize? Because we’re more than just a job!

At voize,
we believe the greatest gift to frontline workers is time - time to care, connect, and be present.
Today, that time is lost to
busywork and complex systems
that pull them away from what matters most:
people
.

Our vision is to change that by building AI companions
that seamlessly take over digital workflows. We don’t replace humans with technology -
we amplify their impact.
Our mission is backed with a
$50M Series A funding led by Balderton Capital, with support from HV Capital, Y Combinator and other leading VCs. Today, 2,000+ facilities trust voize, and over 200,000 users rely on our AI companion
to ease their daily workload.

As a dynamic team, we combine first-in-class technology with meaningful social impact.
And now, we’re looking for you to join us on this mission!
💡 Your Mission: Turning Machine Learning capabilities into real user value

voize is more than a scribe and far more than another documentation tool! We're defining & setting the benchmark what AI in healthcare actually looks like. As
Product Manager - ML
, you’ll sit at the core of our product, you’ll translate customer, Sales, and Customer Success signal into the ML product roadmap across our markets - defining what “better” means for our models, aligning stakeholders on priorities, and driving a shipping cadence that turns ML capability into measurable user value.

🚀 Your Daily Business - No two days are alike

• Own the ML product roadmap for core speech and NLP across our markets, turning field signal into clear priorities
• Run a structured customer-feedback → ML loop with Sales, Customer Support and Customer Success, and direct customer conversations (and close the loop back to the field)
• Define and maintain the model evaluation framework that gates releases (metrics, slices, thresholds, regression bar)
• Partner closely with Speech & NLP Leads as your roadmap counterparts to ship improvements to production
• Improve the product analytics and ML data integration: close the gaps between what we want to achieve, what we can measure in production, and what we train our models on.
• Own ML product KPIs (e.g., no-edit rate, documentation rate, edit rate per slot, transcription accuracy) and drive measurable improvements.

🤝 Your Skillset - What you bring to the table

• Extensive Experience in Product Management, background in speech recognition, NLP, or LLM-based product is a plus.
• Strong technical fluency in ML: evaluation, training data, latency/quality trade-offs, and release mechanics (you don’t need to train models, but you do need to deeply understand them)
• Track record of shipping ML-powered products to production with measurable user impact
• Hands-on experience designing and operationalizing ML model evaluation frameworks
• Excellent communication in English; German is required to lead customer and end-user conversations (roughly B2)
• We are open to relocators, but this role is ideally based in Berlin and you’ll work in a hybrid set-up.

🌱 Growing together - what you can expect at voize

• We are a fast-growing startup, that means you will tackle challenges, grow quickly, and make a real impact, giving frontline workers more time for people
• Become a co-creator of our success with a competitive s stock option program
• Generous perks: 30 vacation days plus your birthday off, Germany Transport Ticket, Urban Sports Club, regular company off-sites, and access to learning platforms such as Blinkist and Audible, plus free language courses
• You decide when you work best, that means flexible working hours and a good hybrid set-up, plus dedicated remote working days within the EU.

✨
Ready to talk?
Apply now! 🚀

We look forward to your application and can’t wait to meet you - no matter who you are or what background you have!