AI & ML Researcher

About Vynexa Vynexa is building the future of AI-powered digital experiences the kind that feel magic, not mechanical. We turn chaotic complexity into clean, intuitive interfaces that make users say: Damn that felt smooth. We dont just ship features. We ship experiences that make every other SaaS product look like its running on dial-up. Were here to win aggressively, elegantly, relentlessly. About The Role We aren't just wrapping APIs. We are building the brain. As an AI & ML Researcher at Vynexa, you wont just be fine-tuning pre-trained models youll be pushing the boundaries of whats possible with applied AI. You will read papers that came out yesterday and implement them today. Youll work at the intersection of research and production, ensuring our models aren't just smart, but fast, efficient, and dangerously effective. If your idea of a good time is reading ArXiv papers and optimizing CUDA kernels, this is your home. Responsibilities Read, dissect, and implement state-of-the-art research papers (Transformers, Diffusion, Mamba, etc.).Train and fine-tune custom small language models (SLMs) for specific domain tasks.Optimize inference pipelines to run blazing fast on consumer hardware.Bridge the gap between pure research and shipping product features.Experiment with novel architectures that give Vynexa an unfair advantage. Requirements Deep intuitive understanding of how LLMs and Diffusion models actually work (not just how to call them).Fluency in PyTorch is non-negotiable.Experience optimizing models (Quantization, Pruning, Distillation, ONNX, TensorRT).Ability to read a research paper and implement it in code within days.Strong software engineering fundamentals your research code needs to survive production.A portfolio of projects that proves you can build things that think.Bonus: Youve published papers or have a significant open-source footprint in the AI community. Compensation We take care of you with Free food and some energy drinks in our apartment because training models takes energy. .