I work as a Lead Machine Learning Engineer. I like
P
Proficient in developing complex AI solutions,
architectures, and scalable deployments on Cloud services. Skilled in building
infrastructure, optimizing models, boosting inference speed, and finetuning
models. Also had experience in teaching web development and developing
Android application.
Leading a team of 7 machine learning engineers to develop, deploy, and optimize advanced AI solutions, enhancing product efficiency and performance.
Architected and developed Chatly/Everask, a state-of-the-art Agentic AI solution integrating text-to-image, image understanding, internet search, website interactions, and RAG document parsing with support for GPT-4, GPT-4-Omni, Gemini 1.5 Pro, and Claude Sonnet 3.5 models.
Spearheading the development of music studio solutions within Imagine.art, expanding the platform's creative capabilities.
Pioneered AI video shorts feature in Video Studio, enabling automated short-form video generation from user prompts, revolutionizing content creation.
Improved scalability and reliability of machine learning services, ensuring seamless integration and deployment across platforms.
Changed model infrastructure from a single pod service to load all models on CPU and on request switch to GPU, eliminating the need to run a lot of pods for each service on a single pod.
Built (Saaz) Singing voice cloning product for the company with its deployment.
Developed professional-grade Product Photography AI solution delivering results in under 2 seconds on consumer GPUs.
I led a team of 5 professionals, overseeing the successful delivery of 22 projects deployed on AWS, GCP, and Runpod. I architected comprehensive solutions across Speech, NLP, and Computer Vision domains, often spearheading projects independently due to their complexity and technical demands. Additionally, I mentored and trained 8 team members, fostering a culture of continuous learning and growth.
Served as an ML Engineer/Technical Lead at Kodezi and led the AI team. Developed end-to-end tools for Kodezi and utilized Node.js to implement efficient pipelines for VS Code extension integration.