Kyle Sayers Email & Phone Number

Medford, Ma, Us

Medford, Ma, Us

Boston, Massachusetts, United States

Contribute to open source machine learning inference tools such as vLLM, LLM Compressor, and Compressed Tensors

Boston, Massachusetts, United States

Served as technical project lead to team of 6 to build a real-time drone imagery mapper product- Developed and maintain high performance image capture and tagging software for commercial UAVs- Developed map tiling service which decreases ingestion latency by 99.1% compared to traditional methods- Researched and implemented SLAM algorithms to meet edge compute and bandwidth targets in C++- Led customer meetings and travelled to demonstrate live mapping capabilities to… Show more Served as technical project lead to team of 6 to build a real-time drone imagery mapper product- Developed and maintain high performance image capture and tagging software for commercial UAVs- Developed map tiling service which decreases ingestion latency by 99.1% compared to traditional methods- Researched and implemented SLAM algorithms to meet edge compute and bandwidth targets in C++- Led customer meetings and travelled to demonstrate live mapping capabilities to clientsMy other work included- Training, evaluating, and deploying deep models for foreign objection detection and robot geofencing- Writing SBIR project proposals for machine learning, computer vision related topics- Developing power-efficient, remotely updatable solar monitoring firmware for the ESP32 in C++- Developing a UAV-tailored SLAM algorithm for live zero-shot stitching- Built retrieval augmented question-answering pipeline using local LLMs and embeddings (personal project)- Built a tool to correlate noisy timestamps using the Kuhn-Munkres algorithm for linear assignment Show less

Boston, Massachusetts, United States

Developed and code reviewed open source software for quantized and sparsified networks with pytorch- Implemented and validated transformer pipelines for fast document embedding and retrieval- Built a zero shot transformer pipeline for mnli inference on nlp data without model retraining- Developed an efficient and accurate tool for analyzing onnx model weights and flops with sparsity

Boston, Massachusetts, United States

Optimized deep machine learning models for stomach cancer diagnosis from biopsy imagery- Refactored codebase for readability and correctness- Implemented and evaluated momentum contrastive learning and bagging ensemble techniques- Improved model accuracy by 7% AUC compared to previous methods

Boston, Massachusetts, United States

Designed, developed, and maintained software for 10+ projects over 17 month co-op- Designed and developed a scalable pipeline for inferencing and displaying aerial data for USDA clients.- Trained generalized U-Net model for segmenting vines from limited aerial imagery- Travelled to Colorado military base to demonstrate deep foreign object detection model- Led team to collect data and train object detection model to detect hurricane survivors- Reviewed remote sensing… Show more Designed, developed, and maintained software for 10+ projects over 17 month co-op- Designed and developed a scalable pipeline for inferencing and displaying aerial data for USDA clients.- Trained generalized U-Net model for segmenting vines from limited aerial imagery- Travelled to Colorado military base to demonstrate deep foreign object detection model- Led team to collect data and train object detection model to detect hurricane survivors- Reviewed remote sensing literature to develop a novel approach for classifying distorted tree imagery- Author of novel SBIR proposal to leverage federated learning of SLAM keypoints on limited hardware- Led project to implement drone packet aggregator system to comply with international regulations- Worked with small team to research and develop an automated photogrammetry pipeline- Created physics-based model using AGX Dynamics to demonstrate custom functionality for DoD client- Mentored high school students from underrepresented communities, assigned programming/cv tasks Show less

Palo Alto, California

San Francisco Bay Area

Taught computer science, math, biology, and guitar to underserved students for 3 summers- Awarded Youth Volunteer of the Year award by Center for Volunteer & Nonprofit Leadership

Bachelor'S Degree, Mathematics And Computer Science

Honors: Magna Cum Laude Extended for sophomore-junior year software co-op

High School Diploma