AI Scientist

This role is for one of the Weekday's clientsMin Experience: 0 yearsLocation: India, Navi MumbaiJobType: full-timeRequirements Advanced AI Research Explore and experiment with cutting-edge architectures in LLMs, diffusion models, transformers, multimodal systems, and reinforcement learning. Reproduce top-tier research papers and extend them through original hypotheses and experimentation. Contribute to internal research documentation, benchmarks, and reproducible experiment pipelines. Model Development & Optimization Train, fine-tune, and optimize deep learning models using PyTorch, TensorFlow, or JAX. Work on tokenization design, data augmentation, loss function optimization, and model scaling strategies. Build high-performance training pipelines using GPUs, TPUs, and distributed compute systems. Applied AI & Deployment Translate research concepts into production-ready, high-impact AI models. Develop robust pipelines for large-scale text, vision, structured, or multimodal datasets. Partner with engineering teams to integrate AI models into production environments. Experimentation & Evaluation Design rigorous experiments grounded in scientific and statistical methods. Benchmark models across accuracy, latency, memory efficiency, robustness, and real-world constraints. Conduct detailed ablation studies and communicate findings with clarity and technical depth. Required Qualifications B.Tech from IITs or other Tier-1 engineering institutions. Preferred Qualifications Experience working with LLMs, diffusion models, GANs, or transformer-based systems. Research internships, academic projects, or publications in AI/ML. Experience with GPU acceleration, CUDA, distributed training, or frameworks like Ray. Contributions to open-source machine learning libraries. Familiarity with cloud ML environments (AWS, GCP, Azure). What Makes This Role Exciting Opportunity to work on high-impact, research-grade AI challenges. Access to advanced compute infrastructure and guidance from experienced researchers. Potential to publish research outputs internally or externally. Fast-moving environment built for ambitious, innovation-driven technologists