Hi, I'm Alexander Borochkin!

I am a Machine Learning Engineer specializing in Visual Language Models (VLMs), computer vision, and generative AI applications. My recent work includes building a cross-platform tool (Thyra) that transforms images and videos into 3D reconstructions and dynamic simulations, integrating state-of-the-art models such as SAM, Mask2Former, GPT-4o, Claude, Gemini, and LLaVA for segmentation, labeling, and temporal reasoning.

I bring hands-on experience with AWS cloud services (AWS Certified Solutions Architect), GPU computing, and scalable ML pipelines using Apache Spark, Kafka, and Docker. I have also applied RAG (Retrieval-Augmented Generation) techniques for multilingual Q&A systems and optimized inference on geospatial data.

Previously, I worked in voice assistant development, where I contributed to international patents and deployed multilingual NLP solutions in production. Earlier in my career, I served as an Assistant Professor in Finance, publishing research on financial modeling and teaching international finance, which provided me with a strong mathematical and analytical background.

Outside of work, I stay sharp through competitive programming (400+ LeetCode problems solved) and continuous professional learning, completing certifications from MIT, Stanford, UCSD, and IBM in AI, data science, and advanced mathematics.

I am passionate about applying cutting-edge AI models to real-world problems, with a current focus on advancing Visual Language Models for video and panoptic segmentation.