Senior Engineer / Small Team — Build CHELT (AI Governance Platform) from Existing Traqplan Codebase

Hello.

I was very interested in your project description because it is exactly the type of system I have been working with in recent years: managed, audited, deterministic AI platforms with clear boundaries of responsibility between humans, the system, and AI.

I work with Node.js, PostgreSQL, and complex backend systems, where critical issues include change control, history, versioning, auditing, access rights, and reproducibility of decisions. I also have experience with Angular and refactoring existing codebases without rewriting everything from scratch.

Additionally, I have extensive practical experience in building AI systems in production with the right constraints: AI does not have direct access to the database; it only interprets intent, generates a structured proposal, and all changes go through the backend, with user confirmation, auditing, and logs. In fact, I have repeatedly implemented the pattern “intent → proposal → confirm → apply → audit.”

How I would implement your basic workflow:
… The user formulates a request in natural language, AI transforms it into a structured change proposal (JSON/DTO), the frontend shows a diff preview, the user confirms, the backend applies it through standard service methods, the change is versioned, and the author, time, previous state, and reason are logged. AI never writes directly to the database and has no hidden state.

Your philosophy resonates with me:
registries as contracts,
strict boundaries between interpretation and mutations,
no “silent” changes for the convenience of AI,
complete traceability and manageability.

From relevant cases:
I built an AI manager for Kommo CRM with RAG architecture, action auditing, controlled scenarios, confirmation logic, and handover to the manager. I also worked on several SaaS systems with workflow logic, approval flows, and change history. I have experience in integrations, process orchestration, and building “managed agents” that not only respond but perform actions under the control of the system.

Your stack and approach suit me perfectly: Node, Postgres, Angular, existing codebase, gradual evolution of the product. I am not an advocate of “rewriting everything,” but rather of careful, controlled strengthening of the architecture.

I am interested in getting involved in this project for the long term and developing it as a product, rather than as a one-off development.

If you want, we can start with a technical conversation; I will look at the current architecture of Traqplan and propose a very specific plan for the first phase of the transition to CHELT.

I would be happy to discuss.

Show original

Translate

ready to help you out with this

have huge experience working in full stack

will send you previous work to make sure we match

Hello!
My name is Andrey Martynov, I represent a group of developers – NC-1.
For over five years, we have been creating websites, mobile applications, online stores, ERP/CRM systems, and other e-commerce products. I offer our team to work on the project:
1. Full-Stack Engineer (senior, team lead). Stack: Node.js, PostgreSQL, Angular.
2. Frontend Developer (middle). Stack: Angular, TypeScript, AG Grid, RxJS.
3. DevOps (middle). Stack: Node.js, AWS, Docker, CI/CD, monitoring.
I answer questions:
1. Examples of production systems
- Incident management platform for fintech: Refactored a monolith in Java into microservices on Node.js (NestJS) and PostgreSQL, adding an AI incident classifier in Python (FastAPI/BERT) with full tracing of predictions for auditing.
… - Multitenant SaaS for marketing automation: Built a workflow orchestration platform from scratch with data isolation through PostgreSQL schemas and managed task execution in Kubernetes, similar to embedded engines (n8n).
2. Layer isolation. Deterministic validation and confirmation. Audit and tracking.
3. I have direct experience in building a multitenant SaaS platform with data isolation at the PostgreSQL schema level. For workflow orchestration, we integrated engines (n8n), adding a management layer: configurations were stored in the main database, and each execution was performed on behalf of a specific agent with injected permissions. Critical processes required mandatory approval through a ticket system before execution. The entire cycle of execution and configuration changes was recorded in a central audit log for traceability.
Cases: https://1drv.ms/b/c/b7a0d31a9dae1bc5/IQCpK38gmEvWT6F_Cso40Li-AXAKkSs-J67mCwll-C732pw?e=v4VVF5
Sincerely,
Alexey M.

Show original

Translate

Hello,

We are Spectrium LLP, a technical team working with production-SaaS and AI functionality with strict requirements for control, auditing, and security. We are interested in long-term work on CHELT AI and transitioning from Traqplan.

🔧 Our experience

Node.js + PostgreSQL — service architecture, data contracts, auditing

Working in existing Angular codebases (refactoring without rewriting)
…
Implementing AI functions without direct access of LLM to the database

🧠 Approach to "intention → application"

AI interprets the request → forms a structured proposal → user confirms → backend applies changes via API → everything is logged and versioned. No hidden states and "silent" mutations.

🚀 What we are ready to close

rebranding Traqplan → CHELT AI

Chelt Hub / Governance Vault

CRUD registries through natural language

AI insights with explainability

foundation for managed action agents

Show original

Translate

Hello.I have been working with Node.js for more than 8+ years.I’m ready to cooperate

Hello.
This project concerns the evolution of the existing SaaS to CHELT AI by implementing managed AI workflows of production level over structured registries, rather than rebuilding from scratch or delivering a thin chat layer. I would work directly in your current Angular/Node/Postgres codebase to rebrand and realign the UX around Chelt Hub and Governance Vault, while establishing strict backend boundaries where AI only suggests changes, and all mutations go through versioned, audited APIs with explicit user approval. For CRUD natural language and analytics, I would implement a clear pipeline intent → structured difference → confirmation → application → audit, treating registry schemas as contracts and maintaining full history and traceability. From there, I would lay the groundwork for managed action agents and workflow execution with strong identification, approvals, and logs, designed for clean scaling in UAT, multi-user, and enterprise environments on AWS.

Show original

Translate

Dear Sir.
We are very interested in supporting the transition of Traqplan into CHELT AI and contributing to the foundation of a governed, production-grade AI-assisted platform.

We have strong senior-level experience with Node.js, PostgreSQL, and Angular-based systems, including working inside existing codebases, refactoring safely, and delivering auditable, enterprise-ready features. We have implemented AI-assisted workflows where intent interpretation is strictly separated from deterministic backend execution, with explicit user confirmation, versioning, and full audit trails.

Our approach to “intent → structured proposal → confirm → apply → audit” follows clear boundaries: AI generates structured, schema-bound proposals only; backend services validate, apply changes via APIs, and log all actions with traceability and history preservation. We are comfortable building multi-tenant SaaS systems and working with governed automation and workflow engines in controlled environments.

We are interested in a long-term collaboration beyond the initial phase and would be happy to share relevant production examples and discuss architecture decisions in more detail.

… Sincerely

Hi there!

This is exactly the kind of product where careful structure matters more than speed, and I am very comfortable working inside an existing codebase and reshaping it without breaking what already works. I have spent years building and refactoring Node and Postgres systems where auditability, permissions, and long lived data contracts were treated as first class concerns, not add ons.

What I like about CHELT is the clear separation between intent and execution. I have implemented similar flows where AI never touches the database directly and instead produces a clean, typed proposal that moves through confirmation, service level validation, application, and full audit logging. That pattern scales well, keeps regulators calm, and makes engineers sleep better at night. Angular systems with dense grids and complex state are also familiar to me, including working around AG Grid performance and schema evolution.

One idea I would bring early is a proposal ledger layer that lives between AI and services. Every intent becomes a signed, replayable object with before and after snapshots and reasoning attached. It gives you explainability, version history, and future rollback or review without adding friction for users.

You can see related production work here.
… https://storiesforkids.ai
https://oscarstories.com

Thank you!

I am interested in your project of transitioning to CHELT AI and can take on both the initial phase and further development of the platform. I have significant experience in Node.js, PostgreSQL, Angular, as well as in implementing AI features into real products with strict requirements for security, auditing, and change control.

🔹 What I can offer for the first phase:
1. Transition from Traqplan → CHELT AI
• Rebranding and updating UX for the new CHELT model
• Adding a dark theme and integrating Chelt Hub and Chelt Governance Vault
• Administrative tools for managing characters and roles
2. CRUD in natural language for registries
• Interpretation of user intent → generation of structured proposals → user confirmation → application via API → log/audit
… • Versioning and change history for all registry schemas
• Secure permission management and control of "silent changes"
3. Analytics and insights
• Generation of explanatory and traceable insights from registry data
• Use of AI without direct access to the database, to comply with security principles
4. Foundations for automated agents
• Action module with approval control, logging, error handling, and retries
• Integration with workflow orchestration (e.g., n8n)

🔹 My experience relevant to this project:
• Development and refactoring of multi-tenant SaaS systems on Node.js / Angular / PostgreSQL
• Implementation of AI for generating proposals, with full auditing and change control
• Realization of "intent → structured proposal → confirm → apply → audit" schemes in production systems
• Working with multi-tenant platforms where critical access control, versioning, and change history are essential

I can quickly get up to speed with the existing codebase, create an initial production-quality foundation, and then maintain and iterate the system alongside your team.

I am ready to discuss the details, assess the scope of the first block, and propose an action plan and timelines.

Show original

Translate