Machine Learning Consultant and Freelance

Cricut Inc, Utah, US, Machine Learning Consulting and Freelance

Max Planck Society

Trainer and Advisor in incubation program for Scientists powered by Max Planck Innovation and Max Planck Society.

Advisor Max-Plank Incubator

Beautiful Destinations

- Make or Buy: benchmark and productionize a computer vision tagging solution for videos and images
- Implement: quickly design and implement several machine learning and data-driven applications including the backend: a video recommender, video similarity search with embeddings, a chat bot, NLP location extraction, a flight price predictor
- AI and data architecture: design intermediate AI outputs like vision tags (e.g. "beach","snorkeling") instead of only end-2-end models. The intermediate outputs enable new features in the product and can be used for marketing

Stack:  Large Language Models (LLMs), GPT-4, Recommender Systems, Google Cloud Platform GCP, Python, Sklearn, Numpy, Javascript, Postgres, Terraform, Docker

Head of AI

Taktile

- NLP financial transaction tagging prediction
- underwriting risk scoring implementation
- reviews and architectural consulting for multiple projects
- python fastapi development

Stack: Python, FastAPI, Sklearn, BigQuery

Freelance Machine Learning Development + Consulting

WhereIsMyTransport

Predict traffic probems and their location from tweets for emerging markets
- developed 2 production ready NLP, state of the art deep learning models that predict cause and effect of a traffic related tweets with good recall and precision, give guidance on future labeling efforts to improve performance
- develop a custom NLP + google maps API pipeline to determine the location of the traffic incident in the tweet. Achieved median error of 60m to the actual location

Stack: Google Cloud Platform, FastAI, Jupyter Notebooks, BigQuery, GoogleMaps, HuggingFace, Poetry

Twitter Traffic Report NLP Solution- Freelance Machine Learning Development

DKB | Deutsche Kreditbank AG

- prototype and implement a service to predict rejected loan applications. Achieved >50% recall at 97% precision, leading to a 5-figure monthly cost reduction
- prototype communal loan price optimization and give guidance on future data collection
- competitive pricing analysis for private loans with machine learning, uncovering multiple big leavers for improved profitability
- ML Ops consultation for all products
- Machine Learning Product Design and Best Practices (Business Case Checklist, review Proposals, Business Case Scouting, ...)

- stack: python, fastapi, pandas, jupyter, sklearn, LightGBM, fastai, docker, grafana, prometheus

Machine Learning Lead Engineer

Zalando SE

• developed and productionized deep learning model and data pipeline for a real-time personalized catalog sorting. The model adapts in real-time to the latest user clicks and creates an article ranking in 6 ms (p99). Users now find more relevant items and need less time to find them. Revenue after returns improved by 2%.
• color based cross-sell recommendations
• design and maintain data pipelines (spark, airflow)
• ops-tasks: Deployment, CI-CD, alerting, maintenance of exiting machine learning products
• hiring engineers and research engineers to create high-performing teams

Senior Research Engineer

• prototype, train, evaluate and productionize algorithms for personalization projects as microservices. Algorithms range from simple collaborative filtering to deep learning in production using Tensorflow.
• design and maintain data pipelines (spark, airflow)
• developed and piloted novel quality monitoring for recommender systems presented at Recsys. It caught multiple permanent quality issues that degraded ROI >15% 
• ops-tasks: Deployment, CI-CD, alerting, maintenance of exiting machine learning products

Strong interest in operational data science (data dev ops) and data quality monitoring.

Research Engineer

• various algorithmic contributions (newsletter recommendations, collaborative filtering, bandit algorithms, ...)
• design and maintain data pipelines (spark, airflow)
• design and maintain an AB/Testing Evaluation Framework in Hadoop

Data Scientist

Quantitativer Analyst in Data Intelligence Department

T-Systems International GmbH

Java development, Upgrade of SOA Message Bus

System Engineer

Siemens

Recurrent Neural State Estimation in Domains with Long-Term Delays

Applied Research, Masterthesis 

I am a pragmatic freelancer and machine learning consultant that likes to solve business problems end-to-end and make machine learning or a simple, fast heuristic work in the real world

Taplio

Lina Weichbrodt's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Lina Weichbrodt's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Lina Weichbrodt's Linkedin Analytics