You will get your codebase set up for Claude Code or Cursor — rules, memory, context
Project details
Your AI agent re-reads half the repository every session, misses your conventions, and re-asks the same questions. That is not the model being weak — it is missing context engineering.
I set your repository up so the agent starts every session already knowing where things live, what the rules are, and what was decided before. I do this daily on my own multi-repo workspace and on client codebases.
What I do:
1. Audit how agents currently fail on your repo — wrong files, ignored conventions, wasted tokens.
2. Write the instruction layer: AGENTS.md / CLAUDE.md, per-directory rules, navigation indexes.
3. Wire the tooling: custom commands, MCP servers, hooks, review checklists.
4. Add a persistent memory layer so decisions survive across sessions.
5. Hand over with a written report on what changed and why.
FAQ
Q: Which tools? A: Claude Code, Cursor, Cline, Windsurf, Codex — the layer is tool-agnostic.
Q: Any language? A: TypeScript, Rust, Python, Go, Solidity.
Q: Do you need repo access? A: Preferred, but I can work from a structure dump under NDA.
Q: What is the measurable result? A: Fewer tokens per task, fewer wrong-file edits, no repeated context rebuilds.
I set your repository up so the agent starts every session already knowing where things live, what the rules are, and what was decided before. I do this daily on my own multi-repo workspace and on client codebases.
What I do:
1. Audit how agents currently fail on your repo — wrong files, ignored conventions, wasted tokens.
2. Write the instruction layer: AGENTS.md / CLAUDE.md, per-directory rules, navigation indexes.
3. Wire the tooling: custom commands, MCP servers, hooks, review checklists.
4. Add a persistent memory layer so decisions survive across sessions.
5. Hand over with a written report on what changed and why.
FAQ
Q: Which tools? A: Claude Code, Cursor, Cline, Windsurf, Codex — the layer is tool-agnostic.
Q: Any language? A: TypeScript, Rust, Python, Go, Solidity.
Q: Do you need repo access? A: Preferred, but I can work from a structure dump under NDA.
Q: What is the measurable result? A: Fewer tokens per task, fewer wrong-file edits, no repeated context rebuilds.
AI Development Type
Knowledge Representation, Model Tuning, Recommendation System, Software MaintenanceWhat's included
| Service Tiers |
Starter
$450
|
Standard
$1,200
|
Advanced
$2,800
|
|---|---|---|---|
| Delivery Time | 3 days | 7 days | 14 days |
Number of Revisions | 1 | 2 | 3 |
AI Model Integration | |||
Detailed Code Comments | |||
Knowledge Graph | |||
Model Documentation | |||
Ontology | |||
Source Code | |||
Taxonomy |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$250 - $1,550
Second repository above the tier
(+ 1 Day)
+$500
Team training 90 minutes
(+ 1 Day)
+$350Frequently asked questions
4 reviews
(4)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
JL
Jue L.
Jan 23, 2025
Presale Widget
Aleksandr is the best developer you'll ever higher. He's incredible at what he does and gets things done quickly and perfectly. You won't regret this hire.
LV
Leonid V.
Sep 28, 2024
Ethereum Node Installation Expert Needed ASAP
I had a great experience working with Alex. He was professional, communicated clearly, and kept me updated throughout the project. He delivered the work on time and exceeded my expectations with the quality and attention to detail. I’m really pleased with the results and would be happy to work with him again in the future.
MJ
Mariusz J.
Sep 24, 2024
Meme Coin Development. Smart Contracts: Presale...
Everything prepared, great communication, very fast work, excellent documentation, well explained and advised in everythong.
MJ
Mariusz J.
Sep 6, 2024
Consultation on smart contracts, Upgradeable .
Great communication, excellent work, everything in time, with excellent quality.
About Aleksandr
AI agents infra / MCP / RAG / multi-agent memory / orchestration
100%
Job Success
Abu Dhabi, United Arab Emirates - 5:53 am local time
I architect and ship production systems: websites, platforms, Telegram bots, integrations. 13 years of engineering, 50+ projects shipped, clients in 12+ countries.
What's different is how I build now. I don't "use ChatGPT": I orchestrate a fleet of AI agents across multiple models and build their tooling myself - RAG over the codebase, memory of past decisions, architecture-aware navigation. My agents don't guess, and don't break what already works.
What I build for clients, I built for myself first and run on it daily. For you that's startup speed at production quality.
• AI agents - agentic systems, orchestration, MCP integrations
• RAG - semantic search over your data, code, documentation
• B2B SaaS - architecture, turnkey builds, CTO-as-a-service
• Automation - Telegram bots, mini apps, internal tooling
• Full-stack - TypeScript, Next.js, Node.js, Mongo/Postgres
• Web3 (8+ years) - Solidity, DeFi, NFT
Steps for completing your project
After purchasing the project, send requirements so Aleksandr can start the project.
Delivery time starts when Aleksandr receives requirements from you.
Aleksandr works on your project following the steps below.
Revisions may occur after the delivery date.
Discovery call
Discovery call — we map which operations your agents actually need, and which must stay read-only.
I design the tool surface
I design the tool surface: names, arguments, output shape, token budget per response.