Senior Deep Learning Research Engineer

Overview

Research, design and develop advanced deep learning and neural architectures for HackerRank's AI/ML systems across its autonomous interviewer, integrity and evaluation problems.

What you'll do

• Research, design and develop advanced Deep Learning models and novel Neural architectures.
• Explore architectures beyond Transformers, including State Space Models (SSMs), Liquid Neural Networks (LNNs), and Hybrid approaches.
• Design and implement Agentic AI systems, Recursive Reasoning Engines, and tool-augmented architectures.
• Fine-tune models using PEFT (e.g. LoRA, QLoRA etc.) and Knowledge Distillation.
• Optimize model Performance, Inference Efficiency, Latency, and Memory utilization.
• Design rigorous experiments, evaluation frameworks, and benchmarks to validate research hypotheses.
• Translate cutting-edge research into scalable, production-grade AI systems.

What you'll need

• M.Tech./M.S./Ph.D. in AI/ML or related disciplines.
• 3-7 years of experience in Deep Learning, Machine Learning, or Applied AI Research.
• Strong foundations in Linear Algebra, Probability, Statistics, Optimization, and Deep Learning algorithms.
• Proficient in frameworks like PyTorch/JAX, with hands-on model development experience.
• Demonstrated ability to implement, reproduce, and extend research papers, including building neural architectures from scratch.
• Familiarity with distributed computing, GPU acceleration, and scalable ML infrastructure.
• Strong understanding of experimental design, model evaluation, benchmarking, and statistical validation.
• Demonstrated research aptitude through publications, open-source contributions etc.

Nice to have

• Publications at top venues (NeurIPS, ICML, ICLR, ACL) or notable open-source contributions in model architectures or agentic systems.
• Experience with LLM evaluation, fraud or anomaly detection, or building systems where the "ground truth" is still being defined.
• Hands-on work taking models from research into low-latency, production inference, such as quantization, distillation, or serving optimization.

Details

• Hybrid, based in Bangalore, India.