π India | π€ Co-Founder & ML Engineer | π Building end-to-end AI products
Currently Co-Founder & ML Engineer at Singularity Forge, building AI SaaS and designing algorithms in the Bittensor ecosystem. Previously a Senior AI Engineer at Samsung R&D, where I developed core computer vision models (segmentation, relighting) for the Galaxy S23, S24, and S25 series. I love turning research into real, scalable software.
π linkedin.com/in/tezuesh | π square-1111.github.io
- Samsung Galaxy AI: Developed instance segmentation and Stable Diffusion LORA portrait relighting models for Samsung S23, S24, and S25 devices (achieving 97.2 IoU at 60ms inference).
- Decentralized AI: Designing algorithms across multiple subnets in the Bittensor ecosystem for mining operations at Singularity Forge.
- Any-to-Any Multimodal AI: Developed speech-to-speech models (Mini-Omni) and a 60,000+ row diarization pipeline at Omega Labs.
- Publications: Co-authored 3DCoMPaT (Composition of materials on parts of 3D things), published in the ECCV 2022 Oral Track.
- π EarthTracks β Environmental mapping leveraging DINO v3, machine learning, and sensor fusion to identify landmines and create 3D environmental models.
- π€ SalesAgent (RevQA) β AI-powered sales coaching platform analyzing call recordings and roleplay sessions for actionable feedback.
- π¨ LogoKraft β App utilizing AI prompt orchestration and brand automation for dynamic logo creation.
- π Reroom β AI room redesign tool that transforms interior spaces with generative AI.
- π€ Mumble β Voice-to-Text app optimized for decentralized knowledge workflows and natural speech generation.
- π€ ClawBot β Conversational AI chatbot with interactive demo.
- π₯ RoastMyWeb β AI-powered website critique tool.
- βοΈ n8n Control Panel β Control panel for n8n workflow automation.
- π¬ 3D-Point-Cloud-Modeling β PyTorch implementation of Generative Models on ShapeNet. Includes custom implementations of Chamfer Distance and Earth Mover Distance. β 26
- πΌοΈ Apery β Creating artistic images using Heuristic algorithms and VGG-based style transfer.
- π Document-Classification β Document Classification using Graph Convolution Networks.
- π Natural-Language-Generation β NLG experiments.
- π Harry-Potter-RNN β Character-level language model trained on Harry Potter.
- 𧬠Genetic-Algorithm β Genetic algorithm implementations.
- π Presenton β AI presentation generator from text prompts.
- βοΈ ScriptGen β AI script generation platform for video and content.
- πΌ JobFinder β AI job matching application.
- π₯ Veo3Plugin β Video generation plugin.
- π΅ MusicCreator β AI music creation tool.





