Are you a full-fledged hands-on engineer who ships fast, cares about user experience, and wants to turn the capabilities of Agentic + Genetic-AI into products that people actually use? Read on.

## Role (what is this job)

Work directly with the CTO and a small, focused team to build AgentPowered; The marketplace where agent creators and companies create, list, discover and transact with AI agents and agent teams. You will have comprehensive features across the stack, integrate LLM/agent tools, and make the platform fast, reliable, and secure.

## What you will build (scope of work)

– User-facing surfaces: Profiles, Lists, Search/Rate, Proposals → Contracts, Invoices, and Payments.

– Backend/API: Cleaned up REST/JSON (and webhooks), authN/Z, inefficient actions, pagination, and rate limits.

– Workflow + Data: Postgres schemas, queues/schedulers, synchronized pipelines, and telemetry events.

– AI Integrations: Inclusions, retrievals, evaluators, and agent/tool ​​runtimes associated with product flows.

– Guardrail: Cost control, tracking, feature flags, and built-in backtracks.

## How it works – Agent Mindset

– Goal → Plan → Action → Observation → Reflecting episodes; Small, measurable Delta shipping weekly.

– Tool Usage: Pick/sequence tools (LLMs, evaluators, retrievers, external APIs) with written I/O and clear contracts.

– Observability: track every important call; See response time, error rates, token spend, and rater success rates.

– Security and privacy: Least privilege, encrypted secrets, reduced PII, and audit capability by default.

– Biased towards clarity: Simple HTML‑first user experience, simple APIs, and tedious infrastructure reduce risk.

– Communicate clearly: English must be spoken and written at a native level.

## Try these maps

– 5+ years shipping end-to-end user facing web products (frontend + backend + Postgres/SQL).

– Powerful TypeScript/JavaScript, HTML, CSS; Comfort with server-served web.

– API versatility (authN/Z, Idempotency, pagination) and schema design.

– Bonus: Embed/RAG/Ratings; Payments/Markets; Multi-tenant SaaS; Astro/HTMX.

## Your agent’s tool belt

### VS Code extensions

– GitHub Helper, Cline, Follow, etc.

– GitGraph, ESLint, Prettier, etc.

– Reduced the price of the mermaid

### Command Line Interface (CLI) and local tools

-gh (GitHub CLI), only or task (repeatable commands)

– Cloudcode, Google Gemini, OpenAI Codex, Augment, Aider, etc.

– Ollama (Local Models), Promptfoo (Prompt Regression/Evaluation)

– psql/pgcli, Docker/Orbstack, local OpenTelemetry collector

### You are likely to connect MCP servers

– File system, git/github, http/fetch

– Postgres (including vector search), Redis/cache

– microsoft/playwright-mcp (validation/capture)

– Stagnation, linearity, idea

– AgentPowered services customized as tools (listings, contracts, returns)

#Who are we?

TalentGenius is the place to find People + AI Agent solutions. We’re a completely remote Americas-based team building the future of how agents build teams, find work, and how companies find agent solutions. We are deeply committed to creating new business models and new technology solutions that will help usher in the era of agents.