Applied AI Model Architect

State Farm•Dunwoody, GA

5d•Hybrid

About The Position

Being good neighbors – helping people, investing in our communities, and making the world a better place – is who we are at State Farm. It is at the core of how we operate and the reason for our success. We're moving with urgency to become a Next Gen Good Neighbor, with technology that strengthens human connections and enhances the customer experience. As an employee, you'll spend less time navigating systems and more time applying judgment, empathy, and expertise in the moments that matter most. In this position, you should plan to spend time working in the office each week and remotely (with the ability for same-day travel to and from your assigned office location) as part of a hybrid work arrangement. Work arrangements could change over time based on business need. Your manager will share additional information regarding your department's approach and what it means for you. There is a 12-month incumbency period (beginning on the employee's effective date) for any lateral job movements and a 12-month incumbency period for any promotional opportunities, which must be met before the employee may post for other State Farm positions. The incumbency period does not affect the at-will relationship between State Farm and the employee and does not create an employment contract, nor contractual rights. Applicants are required to be eligible to lawfully work in the U.S. immediately; employer will not sponsor applicants for U.S. work authorization (e.g. H-1B visa) for this opportunity

Requirements

3-5 years of experience applying transformer-based neural networks and deep learning to real-world problems
A Master's degree in Computer Science, Data Science, Statistics, Math, Actuarial Science, Information Technology, or a related field is required
Hands-on experience with transformer architectures and techniques such as attention mechanisms, self-attention, embeddings, fine-tuning pretrained language models, and transfer learning
Experience applying transformer-based or deep learning models to high-frequency time series data (e.g., forecasting, anomaly detection, or sequence modeling on densely sampled data)
Experience scaling machine learning solutions in resource-constrained environments — e.g., model compression, distillation, quantization, efficient fine-tuning, or infrastructure/cost optimization
Strong programming skills in Python, with experience in deep learning frameworks such as PyTorch or TensorFlow, and libraries such as Hugging Face Transformers
Experience applying NLP/deep learning techniques to structured and unstructured business data; insurance, financial services, or other regulated-industry experience is a plus
Demonstrated experience mentoring or training other data scientists/engineers in model design and experimentation
Strong communication skills and the ability to translate complex modeling concepts for non-technical business partners
Applicants are required to be eligible to lawfully work in the U.S. immediately; employer will not sponsor applicants for U.S. work authorization (e.g. H-1B visa) for this opportunity

Nice To Haves

insurance, financial services, or other regulated-industry experience is a plus

Responsibilities

Serves as a high-level consultant/researcher on department initiatives involving transformer-based neural networks and related deep learning architectures
Designs, develops, fine-tunes, and deploys transformer models (e.g., attention mechanisms, encoder/decoder and encoder-only architectures, embeddings, transfer learning) for underwriting and pricing use cases, including modeling and forecasting on high-frequency time series data
Designs and implements strategies to scale transformer-based solutions efficiently, optimizing for performance and cost within current compute and resource constraints
Reviews and analyzes structured and unstructured data in innovative ways to refine underwriting, pricing, and risk selection objectives across the enterprise
Identifies new data elements, criteria, or factors — including those derived from text, documents, high-frequency time series, and other unstructured or structured sources — for use in underwriting and pricing
Conducts applied research studies in conjunction with other financial/research areas of the enterprise, staying current on state-of-the-art transformer and LLM research
Mentors and trains other data scientists and modelers on the design, experimentation, and evaluation of transformer-based solutions, helping build internal capability in this space
Consults with business clients on the use of modeling, machine learning, and AI, assessing needs and defining project scope
Participates in external professional and industry committees/studies related to AI/ML and NLP
Applies strong programming skills to build and productionize computer-intensive analysis techniques and model pipelines, with an eye toward scalability and efficient use of resources
Routinely engages in quantitative analysis on non-standard and unique business problems
Uses knowledge of business areas and research skills to address and solve a diverse range of real-world business issues

Benefits

Compensation is based on our standard 38:45-hour work week
Potential starting salary range: $190,000-$220,000
Starting salary will be based on skills, background, and experience
High end of the range limited to applicants with significant relevant experience
Potential yearly incentive pay up to 15% of base salary
Annual raise and bonus
Robust health and wellbeing programs
State Farm pays most of your healthcare premium
Multiple healthcare plan options, including a high deductible plan
All medical plans provide 100% coverage for in-network preventative care
Access to vision, dental, telemedicine, 24/7 mental health professionals
Educational benefits like industry leading training programs
Top-notch tuition assistance programs
Employee resource groups
Mentoring
Fertility/IVF/adoption assistance
College coaching
National discount programs
Interactive monthly financial workshops
Free financial coaching
Generous time off policies
Opportunity to initially earn up to 20 days annually plus parental leave
Paid holidays
Celebration day
Life leave (40 hours/year)
Bereavement leave
Community service/education support days
Matching Gift Program
Good Neighbor Grant Program
Employee Assistance Fund
Free financial advisors
401(k) plan with company contributions of up to 7% of your salary