Anomaly Detection Expert in Berlin

• Developed and deployed microservices in Kotlin and Spring Boot, integrated AWS Secrets Manager to secure credentials and decreased network calls using Spring cache.
• Refactored Kafka consumer using Spring Kafka with semaphore-based backpressure to cap records and keep heap memory stable under spikes; switched to batch upserts to cut down on database invocations; added Testcontainers integration tests for Kafka and database to pave the way for future changes.
• Automated the financial reconciliation workflow in Spring Boot (Kotlin) using Spring Scheduler, transactional boundaries, JPA/Hibernate on MySQL, and Flyway migrations, saving the accounts team 16+ hours per week.
• Designed and dockerized payments end-to-end test framework in Robot (Python) with reusable keyword libraries and profiles; integrated with GitLab CI (JaCoCo XML and HTML reports) to accelerate releases and lift code coverage to 80%.
• Implemented end-to-end observability on Datadog by instrumenting services with Datadog APM, correlating metrics and logs, provisioning dashboards, and creating monitors with burn-rate alerts and anomalies to harden reliability and give stakeholders clear visibility.

• Delivered AI and data science solutions, including LLM-based chatbots and data pipelines, improving operational efficiency.
• Collaborated on product features, achieving measurable impact and maintaining strong client relationships.

• Crypto quant strategy development, automated trade execution, onchain data client (Ethereum / Solana)
• Data and trade architecture development for liquidity provision

• Insurance service provider*

Challenge: Business-critical documents were processed through external OCR providers, with ongoing costs, dependency, and data privacy risks for sensitive insurance data.

Implementation:

• Architecture and production implementation of an on-premise OCR solution with full data ownership
• Methods for recognizing document structures as the basis for automated further processing
• ML-, NLP-, and LLM/VLM-based information extraction, especially from invoices and quotations

Success: Replaced external providers: full data ownership, GDPR-compliant processing, and 75% lower recurring OCR costs per year

Used technologies: Python, Docker, Microservices, FastAPI, PyTorch, Torchvision, MongoDB, MySQL

• Conducting cybersecurity readiness checks based on an in-house assessment methodology
• Analyzing the external attack surface using the Graydaxe EASM platform
• Assessing maturity levels and deriving prioritized recommendations for action

• Defined product vision, roadmap, and subscription-based monetization model.
• Architected multimodel AI orchestration (Gemini + GPT fallback) ensuring reliability and cost efficiency.
• Designed explainable insight engine with confidence scoring and agile antipattern detection.
• Built and deployed full-stack architecture (FastAPI, PostgreSQL, React) with secure authentication and quota governance.
• Tech: Python, FastAPI, PostgreSQL, React, TypeScript, Stripe, Gemini API, OpenAI API.

• Large-scale B2C legal services platform operating across Europe in air passenger rights and in Germany across labor, mobility, and rental law.
• Owned the end-to-end onboarding and lifecycle foundations of a high-traffic customer acquisition funnel across 8 European markets, supporting ~10,000 new users per month and backed by €100k/month in paid acquisition spend.
• Led the architectural migration of the core intake and decision logic from fragmented legacy flows into a scalable workflow engine, increasing reliability, improving operational efficiency, and enabling faster iteration on funnel performance and user experience.
• Partnered closely with engineering, operations, and commercial stakeholders to optimize activation, manage complex edge cases, and ensure the platform could scale sustainably across markets without compromising performance or consistency.
• Introduced foundational product analytics and performance monitoring to increase visibility into activation, drop-offs, and lifecycle behavior, enabling data-driven prioritization and experimentation.

• Designed and implemented a real-time perception pipeline using YOLOv7 on Time-of-Flight (ToF) sensor data, enabling live streaming, inference, and on-frame visualization for passenger detection.
• Fine-tuned and evaluated multiple state-of-the-art monocular depth estimation models for Automatic Passenger Counting (APC), and developed a custom hybrid depth model that improved depth accuracy in challenging scene regions.
• Demonstrated that model-generated depth maps outperform raw sensor depth for APC tasks across several datasets, contributing to measurable reductions in counting error.

• Train and optimize MAVAE/VAE models in PyTorch to detect anomalies in multivariate time-series sensor data.
• Design preprocessing workflows and evaluation pipelines to improve model accuracy and robustness.
• Benchmark MAVAE performance against baseline statistical and deep learning approaches, presenting comparative insights.
• Collaborate with research supervisors to refine hypotheses and translate experimental findings into deployable research outputs.

• Led telecom security testing team & SOC deployments across Tier-1 carriers; reduced critical vulnerabilities by 30%.
• Conducted 5G/O-RAN fuzzing, penetration testing, and vulnerability research (basebands, RAN, core).
• Designed testbeds for protocol fuzzing (AFL++, LibAFL) on 5G stacks.
• delivered client workshops on secure telecom with AI assisted workflows.
• Researched AI-driven SOC and penetration testing (LLMs for log triage, anomaly detection, adversarial monitoring).

• Led AI D2C checkout optimization Agent product strategy; instrumented Langfuse for AI evals (task success, latency, cost), iterated on prompts & routing, & drove adoption via cross-team (sales, mktg. & Cust. Success) enablement & 100-merchant launch event
• Led product discovery & built revenue optimization tools, created dashboards with funnel observability using Grafana to track conversion flows, drop-offs, latency, & errors, revenue up by €22.5K+/m
• Built & deployed 3 automation workflows using Claude Code: daily transaction anomaly detection with auto-ticket creation, weekly RCA analysis, monthly feature collation for leadership townhalls, reduced manual effort by 10+ hours/week
• Launched AI voice agent (demo) for e-commerce order & address confirmation/update workflow, designed multi-turn dialogue flows using Pipecat Framework, achieved 71% call pick rate, 100+ Shopify App Store installs
• Owned e2e product lifecycle for 30+ brand (B2B) integrations, collaborate cross-functional teams, ensured payment processing reliability at critical checkout touchpoints, established SLA framework, RCA cadences & ensured 99% SLA adherence
• Led Agile practices as Scrum Master for team of 12, owned sprint & release planning in Jira, established RCA cadences for transaction discrepancy analysis & observability KPIs with Grafana dashboards and delivered 3 major releases on time

• Engaged stakeholders to understand requirements and define the project scope and success criteria
• Demonstrated adaptability by quickly ramping up in a complex, ambiguous regulatory space
• Authored comprehensive science design, architecture review and final methodology documentation ensuring reproducibility
• Gathered data stored in Amazon Redshift and Amazon S3 using SQL
• Performed exploratory data analysis and feature engineering using Python (matplotlib and seaborn), PySpark and Amazon EMR
• Developed and validated machine learning models to facilitate optimization, time-series forecasting, anomaly detection and classification
• Developed machine learning models using Python libraries such as scikit-learn, numpy and pandas
• Deployed the machine learning model using AWS cloud platform (MLOps), especially AWS SageMaker