Lead Engineer
Enregistrez cette offre et organisez votre recherche
Créez un compte gratuit pour enregistrer des offres d'emploi, créer des alertes et revenir à cette liste depuis votre tableau de bord.
En continuant, vous acceptez nos Conditions d’utilisation & Politique de confidentialité.
About us:
At Pencil, we are driving innovation in advertising technology through our state-of-the-art SaaS product, which harnesses Generative AI to redefine content creation. Our mission is to make AI the default in advertising without replacing creative people. To achieve this, we need to make sure that our technology isn’t just in the hands of big brands - we need to try to help small businesses and creative individuals too.
We’re building the agentic OS for marketing. We aren't just using integrating generative AI; we are building the machine that makes it professional, brand-safe, and scalable. We’re moving beyond simple text-in/text-out interfaces toward complex multi-agent architectures that can handle the nuanced demands of global brands and the agility of small businesses.
The Role: Grow the platform our AI agents run on
Pencil's products are increasingly agentic: customers ask for creative work in plain language, and AI agents plan it, call tools and models, and deliver finished assets. Behind that sits a platform that runs those agents reliably, at scale, for many customers at once.
You will be a hands-on technical lead for that platform. You will shape how it scales, how it fails, and how quickly it gets results to users, and you will lead a small team of engineers who build and run it.
We also build software the way we expect our customers to work: with AI. You will set how your team uses AI coding agents to ship faster without lowering the quality bar.
What you'll own:
Agent execution: the systems that run long, multi-step AI tasks: orchestration, tool calling, multiple model providers, and the safeguards that keep long-running work safe.
Scale and reliability: distributed workers on cloud infrastructure, isolation between customers, and graceful handling of failures and deploys.
Data and state: how work, conversations and files are stored and kept consistent under concurrent use.
Real-time experience: streaming progress to users while work is running.
Cost and observability: metering usage accurately, and the tracing and metrics that show where time and money go.
Technical direction: design docs, architecture decisions and code review, and the bar for reliability and testing.
Team leadership: line-manage and grow a small team. Run planning, reviews and on-call, and help hire as the team grows.
AI-driven engineering: set the team's practices for coding with agents such as Claude Code, Codex and Cursor: the context, instructions and test harnesses that make agent output reliable, and where human review stays mandatory.
Your background:
Backend depth: 7+ years building backend systems in production, with Python as a main language. Go is a plus.
Distributed systems: experience running distributed workers or job systems on Kubernetes or similar, including queues, autoscaling and safe deploys.
Data fundamentals: strong relational database and caching skills, including concurrency and performance under load.
LLMs in production: you've built on LLM APIs in production: tool use, streaming, token budgets, and handling provider errors and rate limits. Experience with agent frameworks is a plus.
AI-native engineering: you already work with AI coding agents every day and know how to harness them well. You can break work down so an agent can do it, give it the right context, check its output against tests and real behaviour, and tell where it saves time and where it quietly adds risk.
Team leadership: a track record of leading small engineering teams through delivery: setting direction, mentoring, running reviews, and line-managing.
Evidence-led debugging: you debug from traces, logs and metrics, and you can explain a root cause before proposing a fix.
You'll thrive here if...
A task that took 40 minutes instead of 4 bothers you until you know why.
You care as much about failing cleanly as about the happy path.
You treat your team's dev workflow, AI agents included, as a system worth engineering.
You'd rather delete a layer than add one.
KPI & Success Measures:
Reliability: fewer AI tasks failing or timing out.
Speed: faster time to first result and total task time.
Cost accuracy: usage charges that match real model and compute costs.
Safe scale: more concurrent work and customers without incidents.
Team: a team that ships predictably and makes good use of AI tooling.
<