Projects
A selection of applied AI projects carried out at InstaDeep, spanning Deep Reinforcement Learning, Large Language Models, and Contextual Bandits across industries from aerospace and maritime to biotech and hospitality.
Industry Projects
Large-Scale Deep RL for Airport Resource Allocation
Optimizing delay and workload distribution for Frankfurt Airport using deep reinforcement learning. Deployed and validated in live operations.
Deep RL for Decarbonizing Naval Transport
Optimizing wind-based propulsion embodied systems through adaptive trimming strategies — reducing emissions while maximizing fuel efficiency.
LLM-Based Instruction-Following Robotic Agent
Fine-tuning a large language model for robotic task planning, enabling natural-language instruction following in complex manipulation scenarios.
Deep RL for Robotic Control
Vision-based manipulation in dynamic aerospace environments — combining perception and control for precise, adaptive robotic operation.
Deep RL for Sheet Metal Laser Cutting
Optimizing material waste and cutting paths during metal object manufacturing using deep reinforcement learning.
Deep RL for Diesel Engine Control
Reducing exhaust gas emissions in fleet transportation by learning adaptive engine control policies through deep RL.
Deep RL for Oligonucleotide Synthesis
Leverage RL and genomics foundational models to optimize biological metrics in oligonucleotide synthesis pipelines.
Contextual Bandits for Dynamic Pricing
Optimizing profit for hotel booking through contextual bandit algorithms, including A/B testing and live deployment.