Omelet Insights
Ideas behind better decisions.
Explore how we research, build, and think about operational intelligence. Browse the stories and papers in each field and choose what to read next.
Research & Methods
Explore papers from Omelet’s research archive: optimization methods, learning and search, and multi-agent systems.
PARCO: Parallel AutoRegressive Models for Multi-Agent Combinatorial Optimization
Parallel autoregressive models for multi-agent combinatorial optimization.
Read overview
LPMARL: Linear Programming-based Task Assignment for Hierarchical Multi-agent Reinforcement Learning
Linear-programming-based task assignment in hierarchical multi-agent reinforcement learning.
Read overview
RRNCO: Towards Real-World Routing with Neural Combinatorial Optimization
Neural combinatorial optimization for routing problems with real-world requirements.
Read overview
Neural Embedded Mixed-Integer Optimization for Location-Routing Problems
Neural models embedded in mixed-integer optimization for location-routing problems.
Read overview
RL4CO: an Extensive Reinforcement Learning for Combinatorial Optimization Benchmark
A reinforcement-learning benchmark and library for combinatorial optimization.
Read overview
Neural Genetic Search in Discrete Spaces
Neural genetic search over discrete solution spaces.
Read overview
CAMP: Collaborative Attention Model with Profiles for Vehicle Routing Problems
Collaborative attention and vehicle profiles for routing.
Read overview
Ant Colony Sampling with GFlowNets for Combinatorial Optimization
GFlowNets and ant colony sampling for combinatorial optimization.
Read overview
Generative Flows on Synthetic Pathway for Drug Design
Generative flows over synthetic pathways for drug design.
Read overview
Genetic Algorithms with Neural Cost Predictor for Solving Hierarchical Vehicle Routing Problems
Neural cost prediction within genetic algorithms for hierarchical routing.
Read overview
RouteFinder: Towards Foundation Models for Vehicle Routing Problems
Foundation models for vehicle routing across problem settings.
Read overview
HiMAP: Learning Heuristics-Informed Policies for Large-Scale Multi-Agent Pathfinding
Heuristics-informed policies for large-scale multi-agent pathfinding.
Read overview
Equity-Transformer: Solving NP-hard Min-max Routing Problems as Sequential Generation with Equity Context
Sequential generation with equity context for min-max routing.
Read overview
A Neural Separation Algorithm for the Rounded Capacity Inequalities
Neural separation of rounded capacity inequalities for vehicle routing.
Read overview
Scroll sideways or use the arrows to explore. Select a card to read more.
Engineering
Go inside the systems we build: TOAST’s routing-agent architecture and the role of OIP in running operational intelligence.
OIP: the foundation for running operational intelligence
How shared identity, permissions, deployment, and observability connect optimization capabilities to business applications.
Read story
TOAST: routing optimization through conversation
How TOAST connects a supervisor, specialist agents, and the routing engine, with implementation notes by Juni Lee and co-authors.
Read story
Scroll sideways or use the arrows to explore. Select a card to read more.
Perspectives
Read Jinkyoo Park’s essays and diagrams on AI, optimization, and what changing technology means for industrial decisions.
If agents both build and use software…
An exploration of agent teams, evidence, human supervision, and problem-solving capability.
Read story
Problem solving is not about being good at computation
The engineering judgment behind the balance between a faithful model and a solvable problem.
Read story
The gap between strategy and execution
Why the structure of a decision matters when moving an AI pilot into daily operations.
Read story
A paradigm shift in optimization: data-driven learning and the role of AI
What learning, reasoning, and computing change about industrial optimization.
Read story
OpenAI’s Strawberry and OaaSIS, Omelet’s optimization AI service
How learned solution generation, improvement search, and feedback can form an optimization loop.
Read story
Scroll sideways or use the arrows to explore. Select a card to read more.
Inside Omelet
Meet the people behind the work: reflections between research and entrepreneurship, team gatherings, and how we learn together.
Ten years of SILAB, three years of Omelet: from research to intelligence in the field
A reflection on prediction, control, improvement, and the work of defining real problems.
Read story
Omelet’s first town hall and second birthday
At Omelet’s second birthday and first town hall, the team shared field experience, agent architecture, and product-development direction.
Read story
Two years of solving real problems
A second-anniversary reflection on combining operations research and machine learning with field experience.
Read story
[Between professor and founder] SILAB’s research direction amid technological change
Between professor and founder: choosing a research direction as AI changes the questions we ask.
Read story
An Omelet workshop, in retrospect
A photo journal from Omelet’s October 2024 team workshop.
Read story[Between professor and founder] The decision to start a company
A founder’s reflection on the complementary work of a researcher and an entrepreneur.
Read story
Scroll sideways or use the arrows to explore. Select a card to read more.
Looking for something specific?
Practical learning guides
Explore problem framing, verification, and reusable learning.
Read the guides