Senior AI Engineer

Newfront

17d•$160,000 - $250,000•Hybrid

About The Position

As a Senior AI Engineer at Newfront, you will work across both our AI platform and the AI products built on top of it. You will design and build the core AI platform — the agent runtime, RAG and document understanding pipelines, connector framework, model routing layer, evaluation and observability backbone, and on-premise model hosting — and you will use that same platform to ship AI products that brokers, underwriters, and operations rely on every day. You will partner closely with product engineering, brokering and operations leaders, and security and compliance partners to make sure both the platform and the products reflect how the business actually works. This position is a salary, exempt, and full-time position. This is a US-remote or hybrid role with the option to work from any of Newfront's office locations.

Requirements

BS, MS or PhD in computer science, or related field, or equivalent work experience
5+ years of professional software engineering experience, with a strong general software development background — building, shipping, and operating production services, not just notebooks or prototypes. We are looking for an engineer first, who has gone deep on AI, rather than someone whose experience is exclusively in ML research.
Solid fundamentals in API design, data modeling, testing, debugging production systems, code review, and collaborating in a team codebase.
Strong programming skills in TypeScript, including experience with Node.js or another TypeScript backend framework in production.
Experience with modern development and deployment practices (e.g., containerization, CI/CD, infrastructure-as-code, production observability).
A track record of leading AI/ML projects end-to-end, including API design, production operations, and long-term maintenance.
Experience designing systems for reliability, cost, and scale in production.
Passion for staying up-to-date with the latest advancements in AI/ML and applying them to real-world problems.
Strong problem-solving skills and the ability to take a pragmatic and efficient approach to tackling challenges.
Excellent collaboration and communication skills, with the ability to partner with product teams and effectively communicate complex technical concepts to non-technical stakeholders.

Nice To Haves

Working knowledge of Python and/or Go — we use Python for ML/AI tooling and Go in parts of our backend, and you will read and occasionally write both.
Experience deploying or leveraging machine learning models and Large Language Models (LLMs) to power business applications at scale.
Hands-on experience building agent frameworks, tool-use runtimes, RAG systems, connector / integration frameworks, or evaluation harnesses for LLM-based applications.
Experience self-hosting or fine-tuning open-weight LLMs (e.g., GPU inference serving with vLLM/TGI/TensorRT-LLM, quantization, LoRA/PEFT, on-prem deployment).
Experience building model gateways or routing layers that span multiple model providers and self-hosted models.
Knowledge of state-of-the-art LLM techniques, models, and vendors, including trade-offs across providers.
Familiarity with LLM and related frameworks, including extracting structured data from unstructured text.
Experience with popular AI/ML libraries and frameworks.
Familiarity with DevOps practices, cloud infrastructure, authorization, authentication, and search infrastructure.
Experience with model risk management, AI governance, or building AI systems in regulated industries (financial services, healthcare, insurance).
Understanding of machine learning essentials and the ability to collaborate effectively with data scientists.

Responsibilities

Build and scale the agent runtime at Newfront — tool use, planning, memory, multi-agent orchestration, human-in-the-loop handoff, and the SDKs product teams use to compose agents for brokering, underwriting, and client service workflows.
Design and own RAG and document understanding pipelines for complex insurance artifacts — submissions, policies, endorsements, loss runs, SOVs — including ingestion, chunking, indexing, retrieval, structured extraction, and grounding.
Build the connector framework that lets AI agents and pipelines reach into the systems brokers actually use (AMS, carrier portals, email, document stores, internal services), with first-class auth, rate-limiting, schema discovery, and auditability.
Establish the evaluation and observability backbone for AI at Newfront: offline eval harnesses, regression suites, hallucination and grounding checks, online quality, latency, and cost telemetry, and the dashboards every team uses to measure AI features against compliance and performance targets.
Own model routing and the model gateway — selecting between hosted frontier models and on-premise / self-hosted models per use case, balancing quality, latency, cost, and data-residency constraints.
Stand up and operate on-premise model hosting where regulatory, contractual, or data-sensitivity requirements make hosted APIs unsuitable: GPU capacity planning, inference serving, quantization and optimization, isolation, and lifecycle management.
Ship AI product features end-to-end on top of the platform — from problem discovery with brokering, underwriting, and operations partners through technical design, implementation, and the user-facing surfaces that make AI capabilities intuitive for non-technical users.
Mentor engineers on production AI systems, review designs, and help establish best practices for building reliable AI in financial services.
Partner with security, privacy, and compliance to bake controls — auth, audit, PII handling, retention, model risk management — into the platform rather than leaving them to each product team to rediscover.