Hi, I'm Kriti 👋🏻

Machine Learning Engineer at Apple, working on the Apple Foundation Models that power Apple Intelligence. Parts of my work are open-source via the tamm library on PyPI. I completed my Masters in Computer Science at the University of Wisconsin-Madison.

Outside of work, you can find me…

Journey

1. 

   Apple

   Machine Learning Engineer, Apple Foundation Models

   Jan 2023 – Present

   • Core contributor to Apple Foundation Models, the key framework behind Apple Intelligence experiences. Enables 60+ product teams (Siri, Messages, Photos, Accessibility) to ship on-device and server-side AI with performant models and infrastructure.
   • Developed and publicly released Apple's custom ML library, tamm, on PyPI, empowering App Store developers to build AI features in their own apps using Apple's on-device foundation model. Strengthens privacy-preserving AI across the App Store ecosystem, which generates over $1.3 trillion in developer billings and sales.
   • Led development of a model publishing library that sits in the critical pathway for shipping on-device and server models to millions of devices globally.
   • Designed and productionized a novel model publication and evaluation system, streamlining pre-deployment validation across on-device and server models.
   • Developed a high-performance, vLLM-based inference engine for serving multi-modal foundation models (150B+ parameters); integrated Ray for auto-scaling, load balancing, and fault tolerance to maximize GPU utilization and minimize downtime.
   • Accelerated AI development by reducing model deployment time from several days of manual work to minutes via single-click execution, saving thousands of engineering hours annually.
2. 

   University of Wisconsin-Madison

   Master of Science in Computer Sciences

   Sep 2021 – Dec 2022

   • Research: Domain Adaptation and cross-lingual transfer for NLP with Prof. Junjie Hu; self-supervised video pre-training with masked autoencoders under Prof. Yin Li.
   • Teaching assistant for Machine Learning, Intro to Artificial Intelligence, and Foundations of Data Science.
   • Coursework: Advanced Deep Learning, Advanced Natural Language Processing, Machine Learning, Deep Learning for Visual Recognition.
3. 

   Apple

   Machine Learning Engineer Intern, Machine Learning Platforms

   May 2022 – Sep 2022

   • Integrated the T5 large language model, then state-of-the-art for text-to-text problems, into Apple's internal ML flow tooling so developers across the company could apply it via a simple YAML config.
4. 

   Apple

   Software Engineer, Global Business Intelligence

   Jul 2019 – Aug 2021

   • Redesigned Apple Retail's big-data analytics pipelines (Teradata to a schema-less Spark + AWS S3 architecture) so compute and storage could scale independently. Achieved a 90% reduction in operational costs and handled 100x traffic spikes during Apple's New Product Introductions without disruption.
   • Engineered recovery automations that reduced system downtime from over 4 hours to under 1 minute, safeguarding global retail data operations.
   • Implemented retrieval-augmented (contextual embeddings) search to improve article retrieval for AppleCare advisors.
5. 

   Amazon

   Machine Learning Engineer Intern, Consumer Behavior Analytics

   Jan 2019 – Jul 2019

   • Partnered with research scientists to productionize a Machine Learning Attribution model that guides the apportionment of Amazon's $10B+ advertising budget across channels.
   • Implemented the snapshot creation and featurization modules to process 1B+ records per day, scaling to handle sales and holiday-season traffic spikes.
   • Optimized the machine learning pipeline, reducing the attribution model's execution time by 75%.
6. 

   BITS Pilani

   Bachelor of Engineering in Computer Science

   Aug 2015 – Jan 2019

   • Coursework: Information Retrieval, Optimization, Artificial Intelligence, Data Mining, Data Structures, Fuzzy Logic, Probability and Statistics, Advanced Calculus & Algorithms.
   • Thesis: Code smell prediction using machine learning.
7. 

   Amazon

   Software Development Engineer Intern, Automated Advertising

   May 2018 – Aug 2018

   • Built a logger plugin to aggregate and write Amazon's ads performance data from Meta, X, Pinterest, and other vendors to AWS S3, enabling metrics reporting and anomaly-detection features.
   • Improved customer trust and reduced ticket volume by bringing transparency to anomaly data through a user-friendly dashboard for marketing managers.