Contact
Book a 30-minute technical diagnosis
Share the business context first. I will help assess whether the AI application is worth building, how to approach it, and where the main risks are.
Telegram @NieErAI Message me on Telegram
AI application delivery
I help teams move RAG systems, support assistants, and agent workflows from demo to delivery.
10 years of software engineering experience, building LLM applications since 2024, focused on practical enterprise implementation.
Experience spans securities, regulatory-related work, healthcare, shipping logistics, and environmental utility scenarios, with public material kept anonymized.
Domain LLM Training Toolchain for Regulated Domains
Teams that already have a data source and need to turn a regulated-domain judgment task into their own in-house model — with every decision explainable and auditable. Typically RegTech (securities-regulation content risk control) or compliance-heavy content moderation, where data can't leave the building and calling a generic API isn't enough.
Production Workflow & System Integration
Teams that have validated RAG or document extraction value and need to integrate into real departmental workflows; teams needing to connect AI into OA, CRM, ERP, knowledge bases, ticketing systems, or WeCom.
AI Module Subcontracting & Technical Partnership
Software companies that have client requirements but lack RAG, document extraction, or workflow delivery experience; system integrators needing AI modules, PoCs, bid demos, or delivery support; SaaS teams looking to add AI capabilities without hiring full-time AI engineers.
AI Scenario Diagnosis & Technical Assessment
Business leaders with AI ideas but unsure if they are worth pursuing; teams with a demo that cannot ship to production; teams uncertain about budget, timeline, key risks, and MVP boundaries.
Enterprise RAG Knowledge Base
Teams with policy docs, product documentation, support FAQs, ticket history, training materials, or internal knowledge bases looking to build internal Q&A or support assistance.
Document Parsing & Information Extraction
Teams dealing with contracts, due diligence files, research reports, announcements, PDF tables, scanned documents, and bid documents that need structured extraction.
Selected work
Evidence-oriented cases, projects, and writing.
ComfyUI Multi-Tenant Content Production Backend
The backend core is complete, covering capability contracts, task execution, credit ledgering, permission boundaries, object migration, path traversal protection, and DAG flows. It is not production-proven yet.
ComfyUI Content Production Backend
A multi-tenant SaaS backend that wraps ComfyUI workflows with permissions, credits, node allowlists, and DAG-based production flow governance.
Pauseable data cleaning and LLM augmentation pipeline
A long-running training-data pipeline core using relational corpus storage, DB cursor state, Redis pause flags, and idempotent upsert writeback.
Turning ComfyUI Into SaaS Is Not About Calling /prompt
A practical engineering note on wrapping ComfyUI with contracts, permissions, credits, DAG workflows, and object-storage safety.