18.09.2026 aktualisiert
100 % verfügbar
AI / LLM Engineer
Wiener Neudorf, Österreich
Weltweit
Computer Science & Finance (University of Zurich / ETH Zurich, left in 2021)Skills
LLMFinanzmodellierungPython
AI & LLM Engineering
Design and deployment of MCP servers with scoped tool lists and audit logs, agent infrastructure with retries, dead-lettering and human-in-the-loop gates, RAG pipelines with verification gates that reject unprovable matches, batch inference at controlled cost (~$3 per 1,000 records), and structured outputs constrained by JSON schema using Anthropic and OpenAI APIs.
Python & Rust Backend Development
Production backends written primarily in Python and Rust, including a live crypto market-making desk with ~107k lines of Rust code and 517 unit tests, featuring exchange reconciliation, kill switches, and pre-registered research standards.
Data Engineering & Large-Scale Pipelines
Resumable multi-phase pipelines handling 10M+ requests, proxy pools and per-host rate limiting, dead-letter queues and idempotent retries, official register and exchange APIs across ten jurisdictions, ClickHouse, SQLite/WAL, pandas, and NumPy.
Web Development
Internal tools and review interfaces built with Next.js (App Router), React, TypeScript, and Tailwind, designed for daily production use rather than demonstration.
Infrastructure & DevOps
Linux-based infrastructure with scheduled and supervised services, watchdogs, automated backup, observability and alerting on drift, and cost accounting per workload.
Growth & Marketing Automation
Multi-channel marketing and paid advertising, market launches in the US, France and Italy, influencer software, automated outbound at revenue scale, enrichment waterfalls with cost-ordered fallback, deliverability, DKIM and mailbox verification, and D2C ecommerce.
TypeScript & SQL
Use of TypeScript across frontend and backend projects, combined with SQL for data querying and pipeline management.
Forecasting & Quantitative Modelling
Power price model that outperformed Montel, the leading commercial vendor, achieving 12% lower price error and +5.09 Brier skill points on price spikes over 7,197 shared hours.
ERP & Business System Integration
Integration of ERP systems such as Pohoda, company email, and SharePoint into AI-driven workflows, enabling automated actions, logging, and prompt-injection-safe connectors.
Design and deployment of MCP servers with scoped tool lists and audit logs, agent infrastructure with retries, dead-lettering and human-in-the-loop gates, RAG pipelines with verification gates that reject unprovable matches, batch inference at controlled cost (~$3 per 1,000 records), and structured outputs constrained by JSON schema using Anthropic and OpenAI APIs.
Python & Rust Backend Development
Production backends written primarily in Python and Rust, including a live crypto market-making desk with ~107k lines of Rust code and 517 unit tests, featuring exchange reconciliation, kill switches, and pre-registered research standards.
Data Engineering & Large-Scale Pipelines
Resumable multi-phase pipelines handling 10M+ requests, proxy pools and per-host rate limiting, dead-letter queues and idempotent retries, official register and exchange APIs across ten jurisdictions, ClickHouse, SQLite/WAL, pandas, and NumPy.
Web Development
Internal tools and review interfaces built with Next.js (App Router), React, TypeScript, and Tailwind, designed for daily production use rather than demonstration.
Infrastructure & DevOps
Linux-based infrastructure with scheduled and supervised services, watchdogs, automated backup, observability and alerting on drift, and cost accounting per workload.
Growth & Marketing Automation
Multi-channel marketing and paid advertising, market launches in the US, France and Italy, influencer software, automated outbound at revenue scale, enrichment waterfalls with cost-ordered fallback, deliverability, DKIM and mailbox verification, and D2C ecommerce.
TypeScript & SQL
Use of TypeScript across frontend and backend projects, combined with SQL for data querying and pipeline management.
Forecasting & Quantitative Modelling
Power price model that outperformed Montel, the leading commercial vendor, achieving 12% lower price error and +5.09 Brier skill points on price spikes over 7,197 shared hours.
ERP & Business System Integration
Integration of ERP systems such as Pohoda, company email, and SharePoint into AI-driven workflows, enabling automated actions, logging, and prompt-injection-safe connectors.
Erfahrung
Berufserfahrung5 - 10 Jahre
Freelancer seit2022
Sprachen
DeutschMutterspracheEnglischverhandlungssicherSlowakischMuttersprache
Projekthistorie
Built a company-wide MCP server for a Slovak design wholesaler and retailer with 32 tools over Pohoda ERP, company email and SharePoint. Built a live route feed and communications layer for a transport SME with 50M+ revenue. Building Konteo, an accounting system with own double-entry ledger and LLM decision engine. Built a power price model beating Montel by 12% lower price error. Built a 10M-request resumable pipeline over the Slovak business register extended to ten European jurisdictions. Developed a live crypto market-making desk in Rust with ~107k lines and 517 unit tests.
Built a creator-marketing platform alone, live with 15 paying customers. Implemented outbound grounding for AI-written messages so every money figure, link and promise must exist in a record or the message is held for a human. One person now manages 70-100 creator relationships a month instead of 14.
Built influencer software from scratch covering creator search, automated outreach, product shipping and post tracking. The software brought in 150k/month in new revenue sustained, and reduced the team running it from four people to two.