CXignals core: The engine that powers signal intelligence - ON-1084

Project type: Innovation
Desired discipline(s): Engineering - computer / electrical, Engineering, Computer science, Mathematical Sciences, Mathematics
Company: Numr Inc.
Project Length: 4 to 6 months
Preferred start date: As soon as possible.
Language requirement: English
Location(s): ON, Canada
No. of positions: 1
Desired education level: Master'sPhDPostdoctoral fellow
Open to applicants registered at an institution outside of Canada: No

About the company: 

Numr is a predictive customer intelligence platform that links customer experience directly to business outcomes. Unlike traditional survey tools, Numr continuously listens to customer signals across channels—structured and unstructured—and transforms them into clear, actionable insights.

At the heart of our solution is the Experience Lake, which unifies emotional, behavioral, and transactional data to map customer journeys in real time. Our AI models identify friction points, cluster customers by intent, and predict satisfaction, churn, and revenue impact—across the entire base, not just those who respond to surveys.

Numr stands apart by delivering not just insights, but decisions. Our system automates outreach, recommends fixes, and pinpoints what to improve and where, helping businesses act before problems escalate.

Fully integrable with CRM and operational systems, Numr enables marketing, CX, and operations teams to drive growth, loyalty, and retention with precision. It’s not just customer feedback—it’s a system of business action.

Describe the project.: 

CXignals Core: The Engine that Powers Signal Intelligence is an innovation initiative to build the intelligent backbone of Numr’s predictive customer experience platform. It aims to develop an agentic, self-learning system that detects, interprets, and acts on real-time customer signals—structured and unstructured—across touchpoints.

Today’s customer experience tools rely heavily on survey responses, which capture feedback from less than 10% of customers. CXignals Core reimagines this model by continuously listening to all customers through CRM events, transactions, call center logs, chat transcripts, and behavioral cues. The system then maps patterns to satisfaction, churn, and revenue outcomes.

The innovation lies in evolving a dynamic CX intelligence engine that clusters behavior, predicts outcomes, and recommends specific actions. It shifts CX from passive reporting to active performance management.

The project will involve:
• Enhancing AI clustering and scoring algorithms using unsupervised and reinforcement learning
• Developing context-aware models that link signals to intent and business impact
• Designing self-updating experience graphs that evolve with new data
• Collaborating with academic partners to validate methodologies and publish applied insights

The techniques will span causal modeling, predictive analytics, Bayesian inference, and NLP. A simulation environment will test model stability and performance before deployment.

The result will be a continuously improving agentic core—able to interpret the “why” behind customer behavior and recommend timely, business-aligned actions at scale. By integrating this into enterprise systems, CXignals Core transforms customer experience from a feedback loop into an engine for predictive, profitable decision-making.

Required expertise/skills: 

The ideal candidate should possess a strong foundation in data science, with proven experience in machine learning, natural language processing (NLP), and predictive modeling. Proficiency in Python is essential, particularly with libraries such as scikit-learn, TensorFlow, PyTorch, pandas, and NumPy.

Expertise in unsupervised learning, reinforcement learning, and Bayesian inference is critical, as the project involves developing self-learning, agentic models for customer behavior clustering and signal-based predictions. Familiarity with causal inference techniques (e.g., DoWhy, CausalNex) is also highly desirable.

Experience in handling large-scale, multi-source datasets—including structured CRM data, unstructured text, and behavioral event logs—is expected. Knowledge of graph-based data structures, experience graphs, or knowledge graphs is a plus.
Candidates should be comfortable designing simulation environments for model evaluation and deploying models in production-grade environments using Docker or Kubernetes.

Soft skills include the ability to work in a fast-paced, collaborative environment with researchers, data engineers, and product teams. Prior experience in publishing research, working with academic or applied research teams, or presenting at conferences is a bonus.
Preferred tools and platforms: Python, Jupyter, SQL, AWS (S3, SageMaker), Git, MLflow, and collaborative tools like Notion or Confluence.