AI Solutions Engineer

AI Solutions Engineer (Internal Automations & AI Agents)

What you’ll do (outcomes)

• Ship working systems: Design, build, and maintain small-to-medium services, scripts, and workflows that solve concrete team needs with clear SLAs.
• Automate everything: Orchestrate cross‑app workflows (n8n/Make/Zapier/Custom) across CRM, sheets, Meta/Instagram APIs, email, SMS/WhatsApp, Slack, and web backends.
• Agentic apps: Build LangChain/LangGraph-based agents and tools for data enrichment, routing, triage, and outreach, implementing evaluation, guardrails, retries, and cost controls.
• Scrape—responsibly: Use headless browsers and official APIs to extract data, respecting ToS/robots/legal; implement rotating proxies, backoff, and anti‑bot patterns when appropriate.
• Connect data: Stand up light ETL/ELT to Postgres/BigQuery; build RAG/search over internal docs and metadata; keep data fresh and observable.
• Reliability & telemetry: Add logging, metrics, alerts, and dashboards; own on‑call for the automations you ship (during business hours).
• Documentation & handoff: Write runbooks, diagrams, and clear READMEs so others can extend what you build.

Sample projects you might own

• Instagram lead intake compliant DM triage: Pull creator profiles, enrich, score, then route to a Messenger/IG inbox with auto‑replies and human‑handoff. Track deliverability and outcomes.
• AI outreach bot: Given a segment, generate personalized first‑touch + follow‑ups, dedupe against CRM, schedule messages, and log replies; A/B evaluate prompts and templates.
• Scrape Clean Sync: Headless browser job (Playwright) to capture public business data, normalize, and sync to Postgres/Sheets with change detection and alerts.
• RAG for Sales/Success: Vectorize pitch docs, briefs, and prior wins; build a chat tool that drafts proposals and answers objections with citations.

Our stack & tools (you don’t need all, but you’ve used many)

• Languages & runtimes: Python, TypeScript/Node.js, SQL.
• LLM & agents: LangChain, LangGraph, OpenAI/Anthropic APIs, function/tool use, RAG (FAISS/Pinecone/Weaviate), LlamaIndex (nice‑to‑have).
• Automation/orchestration: n8n (self‑hosted), Make (formerly Integromat), Zapier, serverless cron/queues (BullMQ/Celery), Prefect/Airflow (for heavier jobs).
• Web & scraping: Playwright/Puppeteer, Requests/BS4, Apify; rotating proxies; captcha solving vendors (only when compliant).
• APIs & channels: Meta Graph + Messenger API for Instagram, WhatsApp Business API, SendGrid/Resend, Slack, Telegram, Twilio.
• Data & storage: Postgres/BigQuery, Redis, S3/Cloud Storage, Elastic/OpenSearch; vector DBs (Pinecone/Weaviate/pgvector).
• Infra & DevEx: Docker, GitHub Actions, Vercel/Fly.io/Cloud Run, Terraform (nice‑to‑have), Sentry/Prometheus/Grafana, Feature flags.

What success looks like (90 days)

• You’ve shipped 3–5 production workflows/tools that teams use weekly.
• Lead time from idea pilot 7 days for small asks; 21 days for medium projects.
• Telemetry in place: logs, metrics, alerts, and dashboards for every workflow.
• Documented SOPs enable others to run/extend your work.

Must‑have qualifications

• 4–7+ years building production automations, internal tools, or platform glue (startups or product teams).
• Strong Python and TypeScript/Node; you can ship a CLI today and a small service tomorrow.
• Real experience with LLM apps (tool use/function calling, evals, prompt versioning, cost/perf tuning).
• Confident with APIs/webhooks, auth (OAuth), rate‑limits, pagination, retries, and idempotency.
• Hands‑on with headless browsers and public‑data scraping patterns; pragmatic about ToS/legal.
• Comfortable with databases & SQL, queues, and basic cloud deploys; you instrument what you ship.
• Excellent product sense and communication; you clarify the job to be done and bias to action.

Nice to have

• WhatsApp Business API, Meta Graph (IG), Slack/Telegram bots.
• Vector search/RAG in production; knowledge graphs (Neo4j).
• Data plumbing: Airbyte/dbt/Great Expectations.

How we work

• Small, senior team; async‑first with crisp specs, quick demos, and written docs.
• We value ownership, clarity, and speed. If it’s blocked, you propose a path and ship.
• We optimize for business impact over model elegance.