About the role
Lilt's Converse product delivers real-time, multi-party translation for high-stakes conversations — from offline, device-based deployments in the field to live meetings on enterprise collaboration platforms. We're expanding Converse to support any number of participants, speaking any combination of languages, joining from a growing set of form factors — with live translation delivered N-way, directly on the meeting stage, and cleared interpreters reviewing only the segments that need a human eye. You will build and operate the backend services that make this work: per-participant audio ingestion, real-time diarization and translation orchestration, and the routing logic that gets a flagged segment in front of the right interpreter and broadcasts their correction back out in real time. This is a ground-floor opportunity to help build a brand-new product surface, on a team whose existing products serve some of the world's largest global brands.
As a Backend Engineer on the Create team, you'll work closely with senior engineers and Product to build the services powering this Converse expansion — the real-time translation pipeline and the meeting-platform integration — taking well-scoped services and features from design through production with support and code review from more senior teammates. You'll help define clean API contracts with the frontend/interpreter-workbench engineers, contribute to the event-driven correction and broadcast system, and grow into owning larger pieces of the backend architecture as the product and team scale.
Location & eligibility: This position requires US citizenship and residence in the United States. Preferred locations are Washington, D.C.; Boston, MA; and Indianapolis, IN.
What You'll Do
Build and maintain backend services for the real-time translation pipeline: audio ingestion, diarization hookups, streaming ASR/MT orchestration, and result fan-out to every configured meeting language
Implement the event-driven routing logic that flags low-confidence segments and routes them to the correct on-call interpreter by language
Build APIs and services backing the bot/agent participant and meeting side panel that deliver translation into the host meeting platform, with guidance from senior engineers
Contribute to the interpreter correction and broadcast system — ensuring corrections are isolated to the correct language and broadcast to all participants in real time
Write and maintain automated tests (unit, integration, e2e) and contribute to CI/CD pipelines (GitHub Actions) for new backend services
Instrument services for observability — logging, metrics, tracing — to support debugging of latency-sensitive, multi-party sessions
Partner with frontend/fullstack engineers on the Create team to define clean API contracts between backend services and the live meeting UI and interpreter workbench
Participate in on-call rotation and help triage production issues as the product moves toward GA
Help generate post-meeting assets: diarized transcripts, exportable formats (TXT/WebVTT/JSON), and per-line verification metadata
Required
2–5 years of professional backend software engineering experience, ideally including production systems that handle real-time or streaming data
Experience building REST API endpoints as well as streaming services with WebSockets, gRPC, message queues, or comparable low-latency/event-driven patterns
Solid experience with Node.js/TypeScript, Java, Python, or a comparable backend language/ecosystem, and the ability to work across more than one when needed
Experience building and consuming REST/GraphQL APIs, including authentication patterns (OAuth, API keys/tokens, service accounts)
Comfort with relational/NoSQL data modeling for session state, transcripts, and configuration data
Solid grasp of testing and delivery practices: CI/CD and automated integration/e2e tests
Ability to take a well-scoped ticket or spec and ship it with minimal oversight, while communicating blockers and timeline risk early
US citizenship and residence in the United States
Strong Plus
Experience with speech/audio pipelines: diarization, streaming ASR, or audio stream handling
Experience integrating with third-party platform APIs, SDKs, or webhooks (Slack, Zoom, or similar collaboration platforms)
Familiarity with localization/TMS workflows: XLIFF, Translation Memory, MT vs. human translation pipelines
Experience with event-driven/pub-sub architectures (Kafka, Pub/Sub, or similar)
Enough frontend/React exposure to reason about API contracts with a real-time, stateful UI
Bonus
Hands-on Microsoft Teams app or bot development specifically (Teams Toolkit, Bot Framework, Azure Communication Services, Microsoft Graph)
Experience working on 0-to-1 products within a larger, established company
Interest or coursework in NLP/ML systems
Experience with observability tooling (e.g., Datadog) in production, latency-sensitive systems