Open Positions

About Ontologic Intelligence

Large language models have impressive intelligence but remain limited in their reasoning capabilities. We believe causal reasoning and active inference will be a cornerstone of artificial intelligence. LLMs get causal questions right when the answer is already in their training data, but often fall apart on novel or complex systems. The industry is converging on the idea that LLMs need a causal substrate. We are building it.

Ontologic Intelligence is building scalable persistent causal reasoning infrastructure: the layer between raw data and AI that encodes cause-and-effect with uncertainty, provenance, and support for dynamic, messy real-world structure. With several provisional patents filed, we're founding a team now.

What you'd work on

• Develop and implement novel methods for causal reasoning over dynamic systems with conflicting evidence
• Design and implement a causal graph compiler that ingests messy real-world data and produces structured, query-ready causal representations
• Advance the integration of persistent causal structure with large language models
• Design rigorous experiments to validate intervention and counterfactual queries over compiled causal graphs
• Contribute to patent-backed research on agentic causal graph reasoning, interpretation-conditioned inference, and Causal World Models
• Publish and present research at top venues in causality, ML, and AI

Required Experience

Strong background (PhD, MSc, or equivalent depth through self-directed work) in Computer Science, Mathematics, Physics, or a related quantitative field. We care about what you've done and how you think, not where you went to school. Ideal backgrounds (any of these):

• Causal inference, structural causal models, causal discovery — you think in DAGs, SCMs, and interventions
• Probabilistic graphical models, Bayesian inference, or information theory — you're comfortable with uncertainty as a first-class object
• Graph systems, compiler infrastructure, or scientific computing
• Deep learning research. You can train, fine-tune, and modify language model architectures, not just call APIs
• Neurosymbolic AI — you've worked at the bridge between symbolic reasoning and neural computation

You've built things. You have a personal site, a side project, a tool someone uses, an open-source contribution, a research background in AI or causality, or a startup attempt.

What this is (and isn't)

This is a founding role. Equity-only until we raise (targeting late 2026 / early 2027). No salary yet. You'd be joining at the earliest stage, with an architecture taking shape, filed pending patents, and a clear thesis.

This is for someone who wants to build something foundational, not someone looking for a comfortable job. If you're finishing a PhD, between positions, or working on something that isn't going anywhere and want to work on a problem that matters, this might be the right moment.

Remote works if you're a strong async communicator and self-starter.

How to apply

Send a note to z@ontologiclabs.com. Include your CV, tell me what you've built and why this problem interests you.

Ontologic Intelligence is building scalable persistent causal reasoning infrastructure: the layer between raw data and AI that encodes cause-and-effect with uncertainty, provenance, and support for dynamic, messy real-world systems. With several provisional patents filed, we're founding a team now. A priority research track is causal-gated decoder-only Transformers where learned causal variables condition information flow through attention, MLP, residual, routing, or latent state pathways.

What you'd work on

• Design and implement causal-gated Transformer variants where learned causal variables influence internal computation
• Experiment with gates over attention heads, MLP channels, residual updates, token routing, latent concepts, or expert pathways
• Train and evaluate decoder-only language models from scratch
• Benchmarks and rigorous ablations for intervention, counterfactual, and causal generalization
• Bridge persistent causal world models with large language models
• Work directly with the founder to co-design how structural causal models and cycle-aware graph reasoning become trainable neural architecture components
• Publish at top ML/causality/NLP venues

Required experience

• Transformer internals (attention, residual streams, MLP blocks, training)
• PyTorch model engineering (custom modules, debugging training, controlled experiments)
• Language model training beyond instruction tuning
• Architecture experimentation (changed model internals, measured the result)
• Gating or routing mechanisms (GLU/SwiGLU, MoE, sparse routing)
• Research taste and engineering discipline

Nice to have

• Mechanistic interpretability, causal representation learning, world models, neurosymbolic AI
• Differentiable structure learning, distributed training, FlashAttention
• CUDA/Triton kernels

A founding research role. You'd own the deep learning architecture side — how causal structure conditions neural computation. Early stage, comfort with ambiguity required.

Remote works if you're a strong async communicator and self-starter.

How to apply

Send a note to z@ontologiclabs.com. Include your CV, tell me what you've built and why this problem interests you.

You know your industry's data problems better than any AI researcher ever will. You've seen where machine learning and agentic AI falls short: where predictions based on correlations aren't enough, where you need to know what causes what, where "what would happen if we changed X, and with what probability" is the question that actually matters.

Ontologic Intelligence is building causal reasoning infrastructure: the layer that lets AI reason over interventional and counterfactual structures. We are looking for someone to own a vertical: who knows the buyers, the pain points, the regulatory landscape, and can turn our technology into something that solves an industry need.

What you'd do

• Identify the specific decision-making bottleneck in your industry where causal reasoning would change the game
• Build the business case: who pays, how much, and why they can't solve this with what exists today
• Be the bridge between domain reality and our research team: you know what the data actually looks like, what compliance demands, what buyers care about
• Represent Ontologic at industry events, open doors we can't open ourselves: to pilot customers, design partners, or investors who understand your vertical
• Own the outcome: this isn't an advisory role, it's founding a vertical inside a company

Who this is for

You've spent years in a specific industry — healthcare, pharma, insurance, logistics, energy, agriculture, manufacturing, autonomous systems, legal, finance — and you understand its data infrastructure, decision-making processes, and where the current tools fail.

You might be:

• A senior data scientist or analytics lead who's frustrated that your models can predict but can't explain or intervene
• A domain consultant or advisor who sees the gap between what AI promises and what it delivers in your field
• A founder or ex-founder who still has the domain know-how and knows what the market actually needs
• A technical product lead who's tired of selling incremental AI improvements and wants to build something with real defensibility

You don't need to know what a structural causal model is. You do need to understand your industry deeply enough to know where "why did this happen?" and "what would happen if?" are million-dollar questions.

What this is

Founding role at Ontologic Intelligence. Equity-only until we raise (targeting late 2026 / early 2027). This is for someone who wants founding equity in exchange for opening a vertical. Remote works.

How to apply

Send an email to z@ontologiclabs.com. Include your CV, tell me your industry, the problem you see, and why causal reasoning matters for it.