Utilize your expertise in machine learning, transformer architectures, foundation models, and telemetry-driven AI to impact millions of ArcGIS users worldwide.

Key Responsibilities

Drive the design, development, experimentation, validation, and deployment of AI models built from proprietary telemetry datasets, including both custom transformer architectures and foundation model adaptations

Develop and fine-tune LoRA and QLoRA adapters for language models and sequence prediction systems

Work closely with DevOps and telemetry platform teams responsible for data ingestion, processing, and training infrastructure

Design and implement evaluation frameworks that measure model quality, calibration, throughput, latency, memory efficiency, and operational cost

Conduct rigorous comparisons between custom-built models and foundation-model-based adapter solutions, providing recommendations on architecture and deployment strategy

Build and maintain automated testing and regression frameworks for both base models and adapter-specific functionality

Define and document adapter contracts, including model configuration requirements, tokenizer expectations, input schemas, output behavior, and deployment assumptions

Collaborate with platform engineering, product management, UX, and MLOps teams to deliver production-ready AI solutions with clearly defined capabilities and performance targets

Define training, validation, and benchmarking datasets to support model development and evaluation

Stay current on state-of-the-art developments in transformer architectures, parameter-efficient fine-tuning techniques, model serving technologies, and telemetry-based predictive systems

Author technical design documents, experiment reports, and best-practice guidance for model development and deployment

Mentor software engineers, data scientists,



and analysts on model training, fine-tuning methodologies, telemetry-driven machine learning, and AI research practices

Collaborate with researchers and developers across Esri throughout the AI research and development lifecycle

Solve and articulate complex technical challenges involving machine learning, predictive modeling, and user experience optimization

Requirements

5+ years of professional software development, machine learning engineering, or data science experience

Strong applied machine learning background and deep understanding of modern transformer architectures

Hands-on experience developing and training custom transformer-based models from scratch

Demonstrated experience fine-tuning small and mid-sized language models using parameter-effective methods such as LoRA and QLoRA

Experience building next-item and next-N prediction systems using telemetry, event, behavioral, or sequence data

Experience with PyTorch, Transformer architectures, Foundation models LoRA and QLoRA fine-tuning techniques Model evaluation and benchmarking methodologies

Strong analytical problem-solving skills and experience conducting research-oriented development

Excellent written and verbal communication skills

Ability to communicate complex technical concepts to both engineering and product leadership audiences

Bachelor’s degree in Computer Science, Data Science, Mathematics, Artificial Intelligence, or a related field





Master’s degree or higher in Computer Science, Data Science, Mathematics, Artificial Intelligence, or a related field

Experience with Esri ArcGIS products and geospatial technologies

Experience with adapter composition techniques, including weighted adapter merging, adapter routing, and mixture-of-adapters architectures

Experience with Sequence modeling and predictive analytics

Experience with ONNX Runtime, LlamaSharp

Experience deploying and serving machine learning models in large-scale production environments

Experience optimizing models for constrained environments, including CPU-only, edge, or low-memory GPU deployments

Familiarity with recommendation systems, ranking systems, and behavioral sequence modeling

Experience with retrieval-augmented generation (RAG) and hybrid AI architectures

Experience with graph databases, graph analytics platforms, and graph-based machine learning techniques

Familiarity with large-scale AI inference systems with performance and cost optimization considerations Esri’s competitive total rewards strategy includes industry-leading health and welfare benefits: medical, dental, vision, basic and supplemental life insurance for employees (and their families), 401(k) and profit-sharing programs, minimum accrual of 80 hours of vacation leave, twelve paid holidays throughout the calendar year, and opportunities for personal and professional growth. Base salary is one component of our total rewards strategy.

Compensation decisions and the base range for this role take into account many factors including but not limited to skill sets;

experience and training; licensure and certifications; and other business and organizational needs.

A reasonable estimate of the base salary range is

📌 Sr. Data Scientist- AI Model Development (Winnipeg)
🏢 NextGenEnergyJobs
📍 Winnipeg