Founding Engineer - AI/ML

Uniflow Technologies Inc.; Toronto; Hybrid; CAD 140,000 - 190,000; Full time

Benefits offered by this job

Meaningful founding-team equity

+ Hybrid in Toronto (2-3 days/week in‑

Job summary

Uniflow is building an Artificial Research Intelligence platform for autonomous research with full provenance and verifiable results. We are a seed-stage startup seeking a Founding Engineer - AI/ML to create AI agents covering hypothesis generation, literature survey, experiment design, execution, verification, and paper drafting.

Your work will span multiple domains, ensuring auditable claim chains and provenance-backed outputs.

Qualifications

• 5+ years in ML/AI with production deployment experience.
• Deep expertise in LLMs - fine-tuning, constrained decoding, prompt engineering, evaluation.
• RAG systems - retrieval architectures, embedding models, re-ranking, corpus management.
• Python - PyTorch/JAX, ML infrastructure (training pipelines, serving, monitoring).

Responsibilities

• Understand the execution substrate, provenance model, and verification pipeline.
• Design the retrieval architecture for literature survey with citation tracking.
• Build initial hypothesis generation pipeline with structured output and validation gates.
• Ship production research agent pipeline integrated with the execution substrate.
• Implement regression verification system for claim validation and provenance traces.

Skills

5+ years ML/AI

LLM expertise

RAG systems

Python (PyTorch/JAX)

Tools

PyTorch

JAX

ML infrastructure

Job description

Toronto, ON (Hybrid) Full-time, Founding Team Reports to: CTO

About Uniflow

Uniflow is building an Artificial Research Intelligence platform - end-to-end autonomous research with full provenance, reproducibility, and proof-backed verification. We're a seed-stage startup with a working execution substrate, verification engine, and pilot studies. The concept is proven; the scale-out needs you.

The Role

As Founding Engineer - AI/ML, you will build the AI agents that drive autonomous research - from hypothesis generation through literature survey, experiment design, execution, verification, and paper drafting.

This isn't about bolting a chatbot onto existing tools. You'll design research agents that survey literature with full provenance, generate hypotheses grounded in verified findings, and produce paper drafts with auditable claim chains. Your models will operate across domains - software engineering, medical research, robotics - with auditable claim chains at every step.

What You'll Own

Research agent pipeline (hypothesis experiment verification paper) - Design and build the multi-stage agent system that orchestrates the full research lifecycle, with human-in-the-loop control at every decision point.

RAG infrastructure for literature survey with provenance tracking - Architect the retrieval system (BM25 + vectors + re-rank) that grounds every claim in citable, versioned sources with full evidence tracking.

Claim validation and regression verification - Build the systems that verify generated claims against source evidence, enforce regression gates, and ensure every citation is proof-backed.

What You'll Do

First 90 Days

• Understand the execution substrate, provenance model, and verification pipeline
• Design the retrieval architecture for literature survey with citation tracking
• Build initial hypothesis generation pipeline with structured output and validation gates

First Year

• Ship production research agent pipeline integrated with the execution substrate
• Implement regression verification system for claim validation
• Build the provenance system that makes every generated claim traceable to source
• Achieve measurable reduction in time-to-paper for pilot study partners

Requirements

Must Have

• 5+ years in ML/AI with production deployment experience
• Deep expertise in LLMs - fine-tuning, constrained decoding, prompt engineering, evaluation
• RAG systems - retrieval architectures, embedding models, re-ranking, corpus management
• Python - PyTorch/JAX, ML infrastructure (training pipelines, serving, monitoring)

Nice to Have

• Background in deep learning beyond LLMs (sequence models, representation learning)
• Experience with scientific computing or research automation
• Familiarity with academic publishing workflows or citation analysis
• Knowledge of reproducibility and verification in computational research
• Track record of production ML systems with measurable impact

Engineering Principles

We value:

• RFCs for big decisions - Architectural choices are documented and reviewed
• Reproducibility > speed - We optimize for verifiability and provenance
• Provenance + observability as defaults - Every execution path is traceable and reproducible

Why Uniflow

AI-native from day one - Work at the frontier of autonomous research. Your agents don't just assist - they generate verified, reproducible scientific results.

Execution-coupled intelligence - Your models produce outputs that run on our substrate and feed back into learning. Tight loop from generation to execution to verification.

Founding team impact - Shape the AI architecture, agent design, and verification framework. Early equity participation reflects your role in building the company.

Compensation

• Salary: Competitive, based on experience
• Equity: Meaningful founding-team equity package
• Location: Hybrid in Toronto (2-3 days/week in-office)

Uniflow is an equal opportunity employer. We value diversity and are committed to creating an inclusive environment for all team members.