Julian Kleber

Julian M.Kleber

Applied AI Engineer · Founder

Hey there, I am Pulse, Julian's AI assistant. I am here to answer any questions you have about Julian. Just ask me something to start.

I ship production AI systems that hold up under real constraints.

Get in touch

Case Studies

60% less manual processing

Obsevia: Compliance AI for chemical documentation

FastAPINext.jsSupabaseKafkaPolicy Q&A

Ingest EU safety docs, validate them automatically, and answer regulation questions from the same system. Services scale independently and ship via containers.

View details →
30 min to market plan

Majico: Brand and GTM for micro-SaaS founders

Next.jsMulti-agent AISupabaseRedisJob queues

Name, research, and go-to-market plan in one guided session — work that usually takes an agency weeks. Subscriptions and pay-as-you-go credits, deployed on a VPS.

View details →
1000+ downloads

Agentron: Local multi-agent studio

ElectronNext.jsMySQLLocal modelsSandboxing

Open-source app that runs AI agents on your machine — local data, wired tools, secrets kept out of prompts. Plug-and-play agent setups without cloud lock-in.

View details →
Agentic Ableton sessions

Reeldemo: Agentic Ableton production

AI composeOwn synth engineAbleton LiveSound designStem split

Reference and intent in; a full Ableton handoff out — powered by an in-house synth and sound-design stack. In-house DSP and agent stack, not a bolted-on black box.

View details →
Token-efficient AI-native programming language

Li: Systems language for agents and native speed

Native compilerParallel computeSafety checksAgent queriesCI

Compiled native code that stays within 80% of C++ on published benches, with memory and race bugs caught before run. ~36% fewer AI tokens when agents query data (vs writing SQL by hand), with safer query shapes.

View details →
Cross-chain earnings

YieldScope: Earn rewards from exchanges and chains

Next.jsMonadSoliditySupabaseCEX APIs

Binance, OKX, Monad, and LUNC staking rewards in one ledger — not a portfolio tracker or tax suite. Onchain EarningsCheckpoint so a sync window is portable and explorer-verifiable.

View details →

Intro

About

I started in chemistry labs, then computational science and applied AI. I work from evidence and own the results.

Experience

I ship production-ready AI and high-performance applications in different verticals using a variety of tools, architectures, and technologies.

Focus

Production AI you can audit, with GDPR and EU AI Act constraints from day one.

GitHub

Loading projects…

Publications

Loading publications…

Full profile on ORCID

Connect