Senior AI Security Engineer

About Us

Abacus Insights is transforming how data and AI work for health plans. Our mission is simple: make healthcare data usable, so the people responsible for care and cost decisions can act faster, with confidence.

As the AI and Data Platform for Payer Intelligence, we help health plans break down data silos to create a single, trusted data foundation. That foundation powers better decisions—so plans can improve outcomes, reduce waste, and deliver better experiences for members and providers alike. Backed by $100M from top investors, we're tackling big challenges in an industry that's ready for change. Our platform enables GenAI use cases by delivering clean, connected, and reliable healthcare data to support automation, prioritization, and decision workflows—and it's why we are leading the way.

Our innovation begins with people. We are bold, curious, and collaborative—because the best ideas come from working together. We embrace the thoughtful use of AI and automation to drive innovation and efficiency, and we look for individuals who are curious and adaptable—those excited to leverage emerging technologies to enhance how we work—while keeping human insight, connection, and our clients at the center of every decision.

Ready to make an impact? Join us and let's build the future together.

About the Role

AI is the engine behind where Abacus is headed next, and this role works directly alongside the engineers building that future. We are looking for a Senior AI Security Engineer who is deeply technical and passionate about designing, building, and securing AI-enabled applications that support real customers in regulated healthcare environments.

This is not a traditional security review or compliance role. You will partner with Senior AI Engineers, data quality engineers, platform engineers, DevOps, product owners, and customer delivery teams throughout the development lifecycle. You will help design architectures, contribute code and automation, establish reusable security controls, review implementations, and support secure deployment and operation of Abacus Agents, MCP-based solutions, SQL and RAG agents and tools, and other AI-enabled products.

You will also help enable citizen application development by creating secure paved roads for business-built and AI-assisted applications. The emphasis is on practical engineering: helping teams move quickly while ensuring identity, data protection, authorization, observability, privacy, and compliance are built into each solution from the start.

You will also help build the secure foundations for enterprise knowledge services, including graph-backed knowledge stores and MCP-served retrieval tools that give people and AI agents governed, auditable access to company knowledge. This work spans cloud infrastructure, network, identity, and data layers, so that these services are private by default and secure by design.

This role requires working US Eastern Time (ET) hours, with core hours of 9:00 AM to 5:00 PM ET, Monday through Friday. This is a remote role based in Pune.

Your day to day

Serve as the senior security engineering partner for AI product initiatives, working side by side with Senior AI Engineers and technical leads from concept and architecture through implementation, deployment, monitoring, and production support.

Partner with product managers, tech leads, customer delivery teams, and stakeholders to translate roadmap and customer priorities into secure technical solutions, and translate AI security considerations into practical engineering decisions.

Design, build, and secure AI-enabled applications and reusable platform capabilities, including Abacus Agents, agentic workflows, MCP servers and tools, SQL and RAG agents, model integrations, retrieval services, and customer-facing AI experiences.

Contribute production-quality Python, security automation, reference implementations, tests, policy-as-code, and integration components that help engineering teams adopt secure patterns without slowing delivery.

Define and implement security patterns for identity, authentication, authorization, least-privilege tool access, secrets management, tenant isolation, data classification, encryption, audit logging, telemetry, and human approval controls.

Conduct collaborative architecture reviews and threat modeling for AI applications, including risks related to prompt injection, excessive agency, insecure tool use, sensitive-data disclosure, model and supply-chain dependencies, retrieval poisoning, and unsafe output handling.

Build security requirements and automated controls into GitLab CI/CD, code review, testing, deployment, and release processes, including SAST, SCA, secrets detection, configuration validation, and documented risk acceptance.

Partner with AI Engineering on model and application evaluation, logging, telemetry, and monitoring so security, privacy, reliability, and misuse scenarios are tested alongside functional quality.

Review and support secure implementation of MCP servers, APIs, AI tools, and data connections by evaluating permissions, OAuth and API scopes, token handling, data flows, retention and training terms, vendor posture, and potential blast radius.

Design and build private-by-default cloud foundations for AI knowledge services, including dedicated multi-account environments, private connectivity (VPC endpoint services and transit gateways), firewalled and allowlisted egress, per-account encryption keys, and Terraform-based infrastructure with least-privilege CI/CD deploy roles.

Secure knowledge graph and retrieval platforms, including Neo4j or equivalent graph and vector stores, by designing access tiers, enforcing authorization by caller group, validating backup and restore, and ensuring agents reach data only through governed MCP tools and never through raw database access.

Build and review MCP servers and MCP gateway integrations, including remote transport security, gateway-to-server authentication, identity propagation and its limits, tool-level access policy, and complete audit logging of every call.

Secure ingestion and embedding pipelines that pull from systems such as Confluence, Jira, GitLab, and SharePoint, using read-only service accounts, secrets scanning before content is indexed, rejection of PHI and other regulated data, idempotent re-ingestion, and controlled access to embedding models.

Create and maintain secure paved-road templates for citizen-developed and AI-assisted applications, and help administer platform controls for Replit Enterprise and similar development environments.

Coach citizen developers through application registration, data classification, code scanning, secrets management, GitLab promotion, logging integration, and production readiness while maintaining a predictable review SLA.

Partner with Privacy, Compliance, and Legal to ensure AI applications handling healthcare or customer data meet the privacy and security requirements that apply in each region where Abacus and its customers operate, such as HIPAA, HITRUST, SOC 2, GovRAMP, ISO 27001, and GDPR, including BAA, data residency, and data-boundary considerations.

Represent Abacus credibly in customer-facing architecture and security conversations, explaining how AI-enabled solutions protect data, constrain agent actions, support auditability, and meet customer requirements.

Balance rapid iteration and customer delivery commitments with sound security engineering judgment, and respond directly when design changes, incidents, or client escalations require hands-on technical support.

What you bring to the team

Minimum 5+ years of experience in software engineering, security engineering, platform engineering, DevSecOps, application security, or a related technical discipline, including hands-on responsibility for production systems supporting real customers.

Demonstrated experience designing, building, deploying, or securi