Machine Learning
sprint cycle
The Plank machine learning lab consists of dedicated teams of data scientists, data engineers and mlops engineers that work with innovative technology companies on their most difficult challenges.
Business problem formulation
- Interviewing stakeholders and documenting research
- Objective formulation of business problem along with key metrics
- Defining successful outcome concretely
Data cleaning and data pipeline engineering
- Consolidation of input data sources and schema
- Data analysis to assess missing data, mistyped data
- Creation of consolidated / clean schema of data
- Creating data pipelines
- Deploying data pipeline and batch jobs to populate data lake
Data Science & Model training
- Creating training, test, validation datasets
- Formulating model experiments and tracking model performance
- Tuning of model hyper-parameters
- Optimization of model performance
ML model coding and deployment
- Implementation of prediction API endpoints
- Creation and deployment of model microservices
- Tracking of model performance
- Setup regular model retraining
- Model versioning
Model AB testing
- Integration with data warehouse and analytics services for continuous model performance evaluation
- Defining model variants and AB testing experiments
- Mapping model performance to key business metrics
Customer Showcase
Achaemenid
Achaemenid is a medical technology company that develops innovative products for non-invasive diagnostics, therapeutic, as well as those for management of patients with mild to severe obstructive sleep apnea. Founded by Reza Radmand, DMD, Achaemenid is focused on patient-centric healthcare-related solutions.
.webp)
junction-ai
Junction AI fine tunes the last mile of the supply chain for successful, predictable, and profitable product merchandising. AI success depends on the right infrastructure, tools & data modeling. Junction AI’s proprietary AI/ML platform is an end-to-end data transformation solution. From ingesting structured & unstructured data, cleansing & preprocessing, modeling & insight delivery we’ve combined everything you need for AI/ML success.
.webp)
Migo
Migo is an embedded lending platform that enables companies to extend credit to consumers and small businesses in their own apps. Migo builds proprietary ML algorithms to assess credit risk using the company’s data then automates credit facilities via cutting-edge cloud infrastructure, simplifying the complex world of lending with a simple API. Raised $50million from TPG, Valor Capital, and Nyca Partners
Engineers we’ve worked with
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut morbi bibendum interdum sit. Tincidunt magna nisi, a amet sit enim dui accumsan. Phasellus in nunc gravida.
Matias Lee
PhD in Computer Science from National University of Cordoba
- ex-Instructor of Machine Learning at National University of Cordoba
- 10 years experience in data science and machine learning
- Chief Data Science Officer in multiple companies
- Publications: https:// scholar.google.com/citations? user=eWYixqsAAAAJ&hl=en
