You will get a custom AI agent built on Claude Agent SDK with your own tools
Rising Talent
Rising Talent
Project details
I build AI agents on Anthropic's Claude Agent SDK — the same loop that runs Claude Code, adapted to your specific job. The agent reads a task, picks a tool, runs it, looks at the result, and decides the next step. It keeps context across turns and recovers when a tool fails. You get a working agent wired to your own data and tools, clean Python code you own, a short spec doc, and a Loom demo of the agent running on your use case. Reference build: a Support Triage Agent that classifies 12 tickets, searches a KB, drafts replies, and escalates critical ones in under 90 seconds.
Programming Languages
Python, TypeScriptWhat's included
| Service Tiers |
Starter
$900
|
Standard
$1,900
|
Advanced
$3,500
|
|---|---|---|---|
| Delivery Time | 7 days | 14 days | 21 days |
Number of Revisions | 1 | 2 | 3 |
Number of Pages | 0 | 1 | 5 |
Design Customization | - | - | - |
Content Upload | - | - | - |
Responsive Design | - | - | |
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$450 - $1,750
Additional Revision
+$50
Additional Page
+$100
Extra tool integration
(+ 2 Days)
+$150
Extra subagent
(+ 4 Days)
+$400
Post-launch support 2 weeks
+$300Frequently asked questions
About Richard
AI Automation Engineer | Claude API | AI Agents | LLM | RAG
Giaveno, Italy - 10:05 pm local time
—> I build that layer: classification, routing, estimation, timed responses — all wired into the systems the specific team already uses.
━━━━━━━━━━
🏗️ 𝗥𝗘𝗖𝗘𝗡𝗧 𝗣𝗥𝗢𝗝𝗘𝗖𝗧𝗦
✅ EstateOps AI — three Claude agents for property managers (50-500 units): Maintenance Estimator classifies tickets and assigns responsibility, Vendor Router matches skill and zone, Tenant Comms runs timed messaging (immediate, T-30min, T+2h, T+24h).
✅ AI Voice Receptionist (home care, multi-state) — Retell AI + booking automation, 24/7 inbound handling with qualification and handoff.
✅ Legal Motion Generator — automated intake, AI field extraction, formatted .docx output. Two hours of manual work reduced to 30 seconds per document.
✅ AI Decision Simulator — 350K data points, RAG over Supabase pgvector, Claude + GPT-4 + Gemini orchestration.
✅ Semantic Image Search — CLIP + FAISS, 560 files, offline, sub-100ms query time.
✅ Scraping pipeline — 787K records collected, cleaned, and structured end-to-end with Playwright and Apify.
━━━━━━━━━━
🤖 𝗪𝗛𝗔𝗧 𝗜 𝗕𝗨𝗜𝗟𝗗
✓ Multi-agent systems with Claude (Sonnet 4.6, Haiku) — Anthropic SDK with prompt caching (up to 90% token cost reduction on repeated context) and extended thinking for multi-step reasoning
✓ Voice agents with Retell and VAPI — inbound qualification, outbound follow-up, CRM handoff
✓ RAG pipelines with Supabase pgvector and semantic retrieval at scale
✓ AI automation that replaces manual ticket triage, document processing, and data entry
✓ Data pipelines for scraping, enrichment, deterministic output (Playwright, Apify)
✓ Full-stack delivery — Next.js, FastAPI, Python, TypeScript, Supabase
━━━━━━━━━━
🎯 𝗪𝗛𝗢 𝗜 𝗪𝗢𝗥𝗞 𝗕𝗘𝗦𝗧 𝗪𝗜𝗧𝗛
Founders, property managers, legal and ops teams, home-care operators, and agencies that already know they have a manual bottleneck and want it solved — not studied. The clearer your pain point, the faster I can ship the fix.
━━━━━━━━━━
📩𝗟𝗘𝗧'𝗦 𝗧𝗔𝗟𝗞
Based in Indonesia (WIB, +6 CET). 40+ hrs/week, overlap with EU and US business hours.
If you have a workflow that takes more time than it should, send me a short description — I will reply with how I would build it and a realistic timeline.
Steps for completing your project
After purchasing the project, send requirements so Richard can start the project.
Delivery time starts when Richard receives requirements from you.
Richard works on your project following the steps below.
Revisions may occur after the delivery date.
Kickoff + scoping call
30-min call to define the agent's one job, the tools it needs (APIs, DBs, files), and its inputs/outputs. I confirm the scope fits your tier and flag anything that should be an add-on.
Agent design doc
Short doc with tool specs, system prompt, memory rules, and error recovery strategy. You review and approve before I write any code, so we agree on what the agent will and will not do.