Mid to Senior Software Engineer, backend / full stack

About the role

AutoBid is our automated government procurement bidding platform. It ingests solicitations from government sources — DLA DIBBS, NASA SEWP, SAM.gov, GovWin, and direct email bid feeds — extracts line items and pricing from the documents attached to them, matches opportunities to each customer's catalog and suppliers, orchestrates RFQ email exchanges with those suppliers, and assembles complete bid responses. An LLM-driven pipeline sits at the center of it, and the bids it produces carry real dollars: an extraction error or a duplicated email is not a cosmetic bug.

This is a backend-first role. The hard problems live in the pipeline: parsing inconsistent government documents (PDF, Excel, scanned images requiring OCR), keeping queue-driven workers healthy under memory pressure, evolving a multi-tenant PostgreSQL schema without downtime, and integrating email providers, government portals, and LLM APIs so they fail loudly and recover cleanly. There is a React frontend, and you will ship features in it, but frontend work is a component of this job, not its center of gravity.

The team is small and the ownership is real. You will own features end to end — schema, service, queue topology, API, and the UI that exposes it — and your decisions will persist. We operate in a defense-focused compliance environment, so auditability, least-privilege access, and disciplined logging are part of the engineering culture, not an afterthought.

What you will do

Bid ingestion and document processing

• Build and harden pipelines that ingest solicitations from government sources with different formats, delivery mechanisms, and reliability characteristics
• Extract structured line items and pricing from messy real-world documents: Excel files with hidden characters, image-only PDFs requiring OCR, and attachments large enough that streaming and bounded concurrency matter
• Integrate LLM calls into the pipeline production-grade: structured outputs, validation against expected schemas, retries with backoff, and isolation so one bad document never takes down a service
• Build detection for silent failures — an extraction that quietly drops a line item is worse than one that crashes

Email automation and integrations

• Develop and maintain the RFQ email automation system: outbound dispatch, thread tracking, inbound quote capture, and price extraction from supplier responses
• Integrate with Microsoft Graph and Gmail via OAuth service accounts, applying least-privilege scopes and handling token refresh, reauthentication, and failed-send reconciliation
• Make send paths idempotent: duplicate scheduled emails and double-sent RFQs are real incidents we have fought and won — keep it that way

Backend services and data

• Build NestJS/TypeScript services that consume from RabbitMQ with correct ack discipline, bounded prefetch, graceful shutdown under SIGTERM, and clean behavior through ECS deploys
• Design and evolve the multi-tenant PostgreSQL data model using expand/migrate/contract sequencing for zero-downtime schema changes, including backfills and defunct-table cleanup
• Enforce tenant isolation at the data and service layer — row-level security, company-scoped queries, and tests that prove cross-tenant leakage cannot happen
• Tune query performance, connection limits, and transaction behavior on RDS PostgreSQL as data volume grows

Application development

• Ship full-stack features when the work calls for it: React/TypeScript grids, admin panel workflows, and configuration UIs on top of the backend you build
• Maintain frontend code quality — typing discipline, dependency hygiene, and test coverage — without it becoming your primary surface

Security, observability, and compliance

• Implement structured logging that is useful and safe: no PII, no secrets, no email bodies in production logs; audit-relevant events mirrored to CloudWatch with actor identity
• Remediate vulnerability findings and penetration-test issues on production services and container images as a standing responsibility, not a fire drill
• Engineer with the discipline of a defense-sector compliance posture: access control as an invariant, audit evidence by default, and explainable systems

Required qualifications

• 5–9 years of professional software engineering experience
• Strong proficiency in TypeScript with production experience building server-side services and APIs in Node.js (NestJS experience is a genuine plus)
• Deep hands-on experience with PostgreSQL: schema design, migrations, indexing, query optimization, and multi-tenant data modeling
• Production experience with message queues (RabbitMQ, SQS, or Kafka): consumer semantics, acknowledgement discipline, backpressure, and failure handling
• Experience integrating third-party APIs with OAuth2 in production, including token lifecycle and error recovery
• Experience with document parsing or data extraction from inconsistent, low-quality sources — or LLM-based extraction in production
• Working proficiency in React sufficient to ship complete features without hand-holding
• Experience with structured logging, observability tooling, and production monitoring
• Ability to work independently with significant ownership, make sound technical decisions, and communicate rationale clearly
• U.S. citizenship required

Preferred qualifications

• Experience with AWS services in production: ECS Fargate, RDS (PostgreSQL), S3, ALB, IAM, CloudWatch
• Experience running containerized workloads through deploys: task lifecycle, health checks, graceful shutdown, and memory management
• Experience integrating LLM APIs (Anthropic, OpenAI, or similar) into production systems with structured outputs and evaluation of extraction quality
• Experience with email systems at scale: Microsoft Graph, Gmail API, SMTP providers, deliverability, and thread tracking
• Experience with browser automation or scraping (Playwright or similar) against systems without APIs
• Experience building test infrastructure: integration test suites, seed data, and CI pipelines
• Experience in regulated or compliance-constrained environments (government, defense, healthcare, financial services)
• Background in government contracting, procurement, or supply chain systems is a genuine plus

What makes this different

The output of this system is a bid a customer submits to the federal government, with a hard deadline and real money attached. Reliability is the feature. The most dangerous bugs are silent ones — a price extracted wrong, an item dropped, an email sent twice — and the engineering culture here reflects that: idempotency, validation, and audit trails are how we think, not boxes we check.

You will work on a small team where your decisions directly shape the product. There is no layer between you and the problem: you will read the failing solicitation document, fix the pipeline, and watch the bid go out. Engineers here own features from the data model through the queue topology through the UI.

The work has a compliance dimension most commercial roles do not. We are a defense-focused company, and the discipline that comes with that — least privilege, audit evidence, careful data handling — makes you a better engineer, not a slower one.

Benefits

• Competitive salary and equity compensation
• Comprehensive health, dental, and vision insurance
• 401(k) with employer match
• Flexible PTO policy
• Professional development budget
• Opportunity to obtain a government security clearance sponsored by the company

Equal opportunity

We are an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or veteran status.

How to apply

Submit your resume and a brief note about a production reliability, integration, or data-extraction problem you have solved. Tell us what was failing, what you built, and what you would do differently. We value substance over polish.