Principal Software Engineer - Agentic Platform
Flagship Pioneering, Inc.
| Company | Flagship Pioneering, Inc. |
| Category | Engineering |
| Location | Cambridge |
| Remote | On-site (inferred) |
| Employment | Not stated |
| Level | Lead |
| Salary | Not stated by the employer |
| Posted | 11 Jun 2026 |
| Last verified | 30 Jul 2026 |
| Source | Employer career page (greenhouse) |
Description
ABOUT PIONEERING INTELLIGENCE
Pioneering Intelligence builds on Flagship Pioneering’s legacy of founding cutting-edge science and computational ventures, harnessing recent advances in AI, machine learning, and data to accelerate fundamental research and create a portfolio of AI-first companies. As part of Flagship’s integrated model of science, entrepreneurship, and capital, it transforms breakthrough ideas into world-changing companies, elevating the AI advances happening across the ecosystem in human health, sustainability, and beyond.
The Role
The Pioneering Intelligence group at Flagship Pioneering is building intel ligent s oftware systems that integrate agentic AI with scientific reasoning to accelerate discovery and decision-making. At the center of this effort is a new autonomous research platform. It supports autonomous AI systems that reason over complex biological problems.
We are seeking a Principal Software Engineer to own the underlying agentic platform: the durable orchestration layer on which long-running, multi-agent, multi-model scientific workflows run reliably.
This is a hands-on architect role for a senior engineer who has built and operated distributed systems at scale and wants to t ake age ntic AI from research to production. You will set the technical foundation and standards for the platform, work shoulder-to-shoulder with AI scientists, engineers, and domain experts, and turn frontier agent patterns into reusable, dependable infrastructure.
Key Responsibilities
Own the agentic platform end to end, from research prototype to production, building the durable orchestration layer for long-running, multi-stage scientific workflows. Architect w orkflow orchestration, evaluation frameworks, and event-driven distributed systems that deliver scale, resilience, and reliability.
Define SDK abstractions and API contracts between model and agent outputs and downstream consumers, decoupling rapid model iteration from product stability and reducing integration friction across teams.
Deliver infrastructure-as-code reference implementations (AWS CDK, Terraform, containerized services) and transform exploratory AI proof-of-concepts into hardened, reusable patterns adopted across product domains.
Own platform cost-efficiency and scalability while preserving the performance and accuracy that our experiments and end users require.
Explore and template agent design and deployment patterns, and set technical standards, code quality, and architectural direction for the platform.
Partner across teams (modeling and agent teams, product) to align on integration points, dependency timelines, and delivery, and mentor engineers and scientists.
Professional Experience & Qualifications
B.S. or M.S. in Computer Science, Engineering, or a related field, with 10+ years of professional software development experience.
Demonstrated track record architecting and maintaining large-scale distributed systems in production at technology companies.
Deep expertise in workflow orchestration and event-driven architecture, including stateful execution, idempotency, durable retries, and replay.
Strong cloud-native engineering on AWS with infrastructure-as-code, containerization, and autoscaling.
Proven ability to take systems from prototype to hardened, observable, operationally excellent production.
Experience designing SDKs, APIs, and data contracts that serve multiple downstream teams.
Excellent communication skills and the ability to build organizational consensus around platform decisions.
Collaborative and empathetic working style: listens to understand the needs of scientists, engineers,
991,236 openings. Erioun finds yours.Scored against your own profile, every hour.Try the radar →