Makarand Parigi
Contact Info
- Website: makarandparigi.com
- Email: mparigi99 [at] gmail [dot] com
- Phone: available in pdf
- Location: New York, NY
Experience
- Nextdoor: Machine Learning Engineer on the Vitality ML team, May 2023 - Nov 2023
- Led the development and deployment of an ML model to evaluate posts and comments for potential community guideline violations. This model served as a critical metric for monitoring violative content on the platform.
- Proactively addressed data quality challenges; proposed and executed plan to use GPT to enrich the training dataset, significantly improving model performance (+0.12 PR-AUC).
- Twitter: Machine Learning Engineer II on the Trust & Safety team, Aug 2021 - Feb 2023
- Led team of 3 developers to improve Abuse Detection model by incorporating media features, resulting in
a 10% increase in model precision and my promotion to mid-level engineer.
- Conducted A/B online experiments, resulting in a 15% increase in remediation volume and addition of 2
tweet languages to our scope.
- Migrated tweet report ranking model to Google Cloud's Dataflow, leading to a 63% increase in report action rate
and a decrease in model training time by approximately 90%.
- Independently created an internal web tool to run our abuse detection model with sample inputs.
- Simplified training data generation with BigQuery and refactored TensorFlow model training code to
reduce engineering time required for training new models.
- University of Michigan Radiology: Research Assistant on Deep Learning for MRI Reconstruction, June 2020 - Dec 2020
- Created novel deep learning architecture in PyTorch to "fill in the gaps" in undersampled brain MRI acquisitions.
My method beats the baseline while using 50% fewer parameters.
- Adapting the U-net for Multi-coil MRI Reconstruction
published as an extended abstract at the International Society for Magnetic Resonance in Medicine (ISMRM) 2021.
- Presented at Neuromatch 3.0, a neuroscience conference.
- University of Michigan Computer Science: Teaching Assistant on {Programming Languages Theory, OOP and Data Structures in C++, Intro to Programming}, Jan 2018 - Apr 2021
- Led development of instructional materials for 1000+ students per semester.
- Hosted exam review sessions, taught discussion sections, delivered lectures, and assisted students in office hours.
Skills
- Languages: Python, SQL, C++, Haskell
- Technologies: Tensorflow/Keras, PyTorch, BigQuery, Apache Beam, Google Cloud (GCP)
Internships
- Twitter: SWE Intern on the JVM team, Summer 2019
- Amazon: SWE Intern on the Growth Engagement team at Amazon Music, Summer 2018
- Designed and implemented AWS architecture so marketing campaigns could send push notifications into the web application.
Used new browser technologies (Service Workers, Push API).
Education
- M.S. Computer Science, University of Michigan. GPA: 3.64, 2021
- B.S.E. Computer Science, University of Michigan. GPA: 3.91 (Summa Cum Laude), 2022