Incode Technologies - Growth Engineer

• Engineering foundation • 4+ years professional software engineering • 1+ year building and shipping LLM-powered applications or agents to production - used daily by people who are not engineers • Strong Python and Node.js/TypeScript in production; React sufficient to build internal UI end to end • At least one internal application you built start to finish: data model, backend, interface, auth, deployment, and the maintenance afterward • Hands-on experience with both Salesforce and HubSpot. We run both, and this role writes against both. We are not asking for admin certification - we are asking for working fluency: standard and custom objects, field-level behavior and validation, record ownership and sharing, REST and bulk APIs, and a clear sense of what breaks downstream when you change something upstream • Sales engagement platform experience - Outreach, Salesloft, or Apollo: sequence and step models, prospect and mailing data structures, and API access patterns • A realistic view of how revenue data degrades in practice: duplicates, stale ownership, unenforced picklists, half-migrated fields - and what that means for anything automated on top of it • Bonus: enrichment and contact data providers (ZoomInfo, Clay, Lusha-class) at the API level rather than the UI level • Databases and data work • Production experience with Postgres or Supabase: schema design, migrations, indexing, query tuning, row-level security • SQL fluency against a warehouse - Redshift, Snowflake, or BigQuery-class - including window functions and incremental models • You have reconciled the same metric across two systems, found the delta, and explained it to a non-technical stakeholder • Automation, end to end • You have designed, built, deployed, monitored, and maintained automations in production - not just built them. You own the alerting and the failure modes for your own work • Webhooks, queues, scheduled jobs, retries, idempotency, and backfills. You have debugged a job that silently stopped firing • Automation platform experience (n8n, Tines, Zapier-class) and the judgment to know when to abandon the platform and write the service instead. We care more about the second half of that sentence • Routing, scoring, and decision logic implemented as code with test coverage, not as a chain of conditional blocks nobody can read • AI and agents • Hands-on with MCP - servers, clients, custom endpoints - or an equivalent tool-calling and agent framework • Production LLM API integration (Anthropic-class): structured output, tool definitions, context management, cost and latency tradeoffs • Prompt and skill design treated as versioned artifacts, with some method for evaluating whether a change helped • Deployment, integration, and security • Cloud deployment ownership: AWS, containerized hosting, CI/CD (GitHub Actions); Terraform or another IaC tool a plus • OAuth flows, token lifecycle and refresh, credential vaults, header auth, and least-privilege scope review on the integrations you build • Experience supporting a multi-region revenue team (NAM, LATAM, EMEA) and the data-residency questions that come with it • Event and conference lead-capture tooling • Notion and Slack API work • Experience enabling non-technical teams on AI tooling - documentation, training, internal adoption • The Work You've Done BeforeCome ready to walk us through two or three things you have built. The strongest candidates we have seen have shipped work like: • The Work You've Done Before • An internal application a non-technical team now uses every day, where you owned it from data model to deployment • A two-way integration between two systems that had no native connector, including how you handled auth, rate limits, and partial failure • An agent or automation that replaced a recurring manual process, with a before-and-after number attached to it • A data reconciliation between a CRM and a warehouse where the two disagreed and you determined which one was right • Something you built, documented, and handed off - that someone else now runs without you • We care more about what you shipped and what broke than about which tools were on the sticker. If your strongest work used a stack we did not list, tell us about it anyway.