AI research scientist
About shoppin’
shoppin’ is an AI - powered visual fashion search engine - if google’s search exhaustiveness and pinterest’s social DNA were to have a baby, it’d be us.
Today, gen-z shopping is super trend and intent-led, where they know exactly what they want to look for. our multi modal search engine allows you to discover fashion with personalised recommendations, search with images and prompts to find that exact look or the closest looking products from across 50k+ brands.
We've raised $1 million pre-seed and are backed by one of india’s leading vc firms.
To know more about shoppin’ refer to: https://www.canva.com/design/DAGZnaBOCTQ/RXLVAkBUSogVR8gWy2gEzA/edit our ig handle: our founders : cto, Utsav Soi ; ceo, Shlok Bhartiya
About the role
We seek a technically rigorous AI Research engineer to advance our core AI systems, with a focus on training, fine-tuning, and optimizing large language models (LLMs). You will bridge cutting-edge research with production-grade engineering, designing high-performance implementations for GPU/TPU clusters and iterating on agentic workflows. This role demands fluency in both theoretical foundations (e.g., transformer architectures, optimization theory) and applied engineering to deploy scalable AI solutions.
Key responsibilities
- Great at maths and algorithms, with a strong foundation in problem-solving.
- Deep understanding of the first principles behind language models and their architectures.
- Design, train, and fine-tune Transformer-based models and LLMs on multi-GPU systems with cutting-edge techniques.
- Lead research in foundational AI areas, including pretraining, RLHF, efficient fine-tuning methods (e.g., LoRA, QLoRA), and model distillation.
- Rapidly prototype and integrate state-of-the-art techniques from AI research papers into production-ready codebases.
- Optimize distributed training pipelines for large-scale LLMs across GPU clusters using frameworks like PyTorch, CUDA, and FSDP.
- Orchestrate agentic workflows, including multi-agent systems, tool integration, and recursive optimization strategies.
- Develop high-performance implementations of deep learning algorithms, including custom kernel development with Triton and CUDA.
- Stay at the forefront of AI research, exploring trends like emergent behavior in LLMs, speculative decoding, and Mixture-of-Experts scaling, while assessing their practical applications.
What we offer:
- competitive salary and esops, along with hackerhouse living: live and work with a gen-z team in a 7bhk house on mg road, gurgaon.
- hands-on experience in shipping world-class products, professional development opportunities, flexible hours, and a collaborative, supportive culture.
join us to revolutionize the future of online shopping!