AI Research Engineer
What is the role?
At Tensorplay, staying at the frontier of AI research isn’t optional — it’s how we deliver value. As our AI Research Engineer, you’ll monitor and evaluate emerging techniques from arXiv and top-tier conferences, run experiments to assess their practical applicability, and implement the best ideas as production features for our clients.
You’ll work closely with our ML and backend engineers to take techniques from paper to production.
About you
You’re fluent in both the language of research papers and the pragmatics of production systems. You can implement a novel attention mechanism from scratch, but you also know when it’s wiser to fine-tune an existing model. You’re excited by advances in reasoning, multimodal models, efficient inference, and agent architectures.
Your Responsibilities
- Continuously monitor AI research across arXiv, NeurIPS, ICLR, ICML, and major labs
- Evaluate and prototype promising new techniques for client use cases
- Run systematic experiments with clear benchmarks and ablation studies
- Implement and adapt state-of-the-art methods for production constraints
- Write internal research notes and client-facing technical reports on AI evaluation results
- Contribute to Tensorplay’s public AI research blog and community presence
Requirements
- MS or PhD in Machine Learning, Computer Science, or related field (or equivalent research experience)
- Strong publication record or hands-on experience implementing published SOTA methods
- Deep understanding of transformer architectures, attention mechanisms, and modern LLMs
- Proficiency in Python and PyTorch; ability to read and implement from papers
- Experience with distributed training and large-scale model experimentation
- Strong written communication skills for research documentation
Nice to Have
- Experience with PEFT methods (LoRA, QLoRA, Adapters) and alignment techniques (RLHF, DPO)
- Familiarity with multimodal models (vision-language, audio-language)
- Knowledge of efficient inference techniques (quantization, speculative decoding, flash attention)
We Offer
- Competitive salary with equity
- Dedicated GPU compute budget for research experiments
- Work on real-world AI problems with meaningful impact
- Collaborative research environment with brilliant engineers