
Applied AI Engineer - Agent
The General Intelligence Company of New York12 months ago
New York, NY, USAMid Level
Base Salary
$250k - $300k/yr
Responsibilities
- Design and implement agent improvements across prompting, tool selection, action planning, memory, safety guardrails, and recovery paths.
- Build offline and online evaluation pipelines, experimentation frameworks, regression suites, behavior tests, and guardrail thresholds.
- Productionize function and tool-calling orchestration, self-reflection, retrieval/RAG, multi-agent handoffs, caching, embeddings, and hallucination-reduction techniques.
- Improve backend orchestration, retries, backoff, idempotency, auditability, memory and context routing, integrations data pipelines, and agent observability.
- Partner with product and infrastructure teams to define success metrics and ship safe iterations.
- Write tested code, document design decisions, and maintain runbooks.
Requirements
- 4+ years of backend engineering experience, preferably with Python.
- Hands-on experience with prompt engineering, function calling, retrieval, embeddings, evaluation design, and shipping LLM features to production.
- Experience building evaluation harnesses and using regression suites, task-success metrics, and cost/runtime tradeoffs to drive improvements.
- Strong distributed-systems fundamentals, including concurrency, reliability, performance, data modeling, and lifecycle management.
- Pragmatic experimentation skills spanning prototyping, measurement, and rollout.
- Excellent debugging and instrumentation skills.
- Experience with agent frameworks, tool orchestration, memory architectures, production RAG, Redis, Postgres/Supabase, queues, event-driven designs, Datadog, or OpenTelemetry is nice to have.
Benefits
- Meaningful equity and comprehensive benefits are offered.
- Flexible work setup is available.
Tech Stack
Categories
About The General Intelligence Company of New York
The General Intelligence Company is building highly-capable ai agents to run entire businesses autonomously.