Sunipun Saikat
AI Engineer & Software Developer
I'm an AI Engineer and software developer, with a strong foundation in mathematics and a passion for solving real-world problems through artificial intelligence. I bring hands-on experience in machine learning, computer vision, and generative AI. Building systems used by millions of users globally.
Generative AI & Diffusion Models
Engineered production-level models for image and video generation, enhancement, and transformation. Extensive experience in text-to-image, text-to-gif, image-to-video, and faceswap pipelines.
Computer Vision Applications
Deployed scalable models for segmentation, classification, denoising, and super-resolution, impacting over 10 million users through commercial apps like Cartoon AI.
Mobile & Multimedia Engineering
From Android internals to OpenGL rendering, built multimedia modules in Java/Kotlin/C++ optimizing performance for devices across the spectrum.
Research & R&D Pipelines
Experienced in training and fine-tuning ML models with PyTorch, Transformers, and VectorDBs, including work on semantic-aware image search and custom image classification pipelines.
Professional Experience
AI Engineer
BrainCraft Ltd, Dhaka, Bangladesh
December 2022 - Present
- Engineered Generative AI models for image creation in Cartoon.ai, propelling it to top-grossing status
- Optimized AI-driven photo enhancement modules, using open-source models, incorporating upscaling, deblurring, denoising, and artifact removal, which improved image quality for over 500K users
- Collaborated intensively in developing and deploying a semantic image segmentation model, elevating user experience and delivering precise segmentation functionality to a user base exceeding 10 million
- Implemented a comprehensive system for text-to-image and text-to-gif generation using stable diffusion models and Python, leveraging advanced machine learning techniques and image processing algorithms
Software Engineer
BrainCraft Ltd, Dhaka, Bangladesh
January 2022 - November 2022
- Planned and implemented an Image to Video making module using OpenGLES and MediaCodec API for the GIFMaker and Add Music to Video android apps, resulting in over 250k downloads on the Google Play Store
- Developed a core video encoder/decoder module using raw C++ libraries and the Android NDK, increasing efficiency on lower-end devices and expanding availability by 22%
- Managed and guided a team of four members in the successful development of two feature-rich BgRemover and Slideshow apps, leveraging Android Core functionalities, MVVM architecture, custom rendering techniques, and advanced video encoding/decoding technologies
Software Engineer
LiiLab, Sylhet, Bangladesh
January 2021 - December 2021
- Architected and Programmed the Intro Maker app, leveraging Canvas and Native Android APIs to create stunning intros. Achieved an impressive milestone of over 100k downloads on the Play Store
- Optimized GkQuiz's rank list functionality using Firebase Realtime Database and Firestore, achieving a remarkable 4x performance boost. Attracted over 250k users on the Play Store
Education
Shahjalal University of Science And Technology
Bachelor's in Mathematics (2015 - 2020)