Open for Founding Engineer / Architect Roles
FFFF
oooo
uuuu
nnnn
dddd
iiii
nnnn
gggg
 
EEEE
nnnn
gggg
iiii
nnnn
eeee
eeee
rrrr
&&&&
SSSS
yyyy
ssss
tttt
eeee
mmmm
ssss
 
AAAA
rrrr
cccc
hhhh
iiii
tttt
eeee
cccc
tttt
AI Architect Portrait
YYYY
uuuu
llll
aaaa
iiii
 
KKKK
aaaa
rrrr
iiii
mmmm
oooo
vvvv
SYS_METRICS
v2.0.1
SYS_01
0+
Shipped Systems
Full-Cycle Dev
SYS_02
0
Active Users
Organic Growth
SYS_03
0%
Tender Screening
Manual Review Reduced
Enterprise RAG
B2B Knowledge Engines
Autonomous Agents
Browser Automation
High-Load B2C Apps
TG Bots & WebApps
High-Agency
MVP to Prod < 2 Weeks
Diverse Engineering
Complex Business Logic
TECH_STACK
8 tools
Python
FastAPI
Docker
Qdrant
LangChain
React
Celery
PostgreSQL
Python
FastAPI
Docker
Qdrant
LangChain
React
Celery
PostgreSQL
Python
FastAPI
Docker
Qdrant
LangChain
React
Celery
PostgreSQL
PROJECTS
selected work
Production

TenderAI Platform

/ 500+ Docs / Day

Enterprise RAG screening 500+ commercial documents, scans, and spreadsheets daily. Reduced manual tender review by 90%.

Heterogeneous DataQdrantOCR Pipeline
View
R&D Tooling / Architecture Concept

Autonomous OSINT & Enrichment Engine

/ Multi-Tenant Architecture

Agentic Data Pipeline. Automated pipelines normalize unstructured data from Telegram, HTML, and APIs with Supabase-style multi-tenancy.

OSINT PipelinesEAV SchemaText-to-SQL
View
In use by law firms

Bankruptcy AI

/ Document Intelligence

AI-assisted case preparation for bankruptcy law firms. Turn credit reports, bank statements, and scans into a source-linked report for specialist review.

Document ParsingFinancial AnalysisPDF / DOCX
View
Delivered integration

AI Head of Sales

/ CRM + 1C + Messengers

Customer intake, a staff copilot, and call analysis connected to CRM and 1C. Approved actions, follow-up reminders, and team summaries in messengers.

AI AgentsCRM / 1CSales Operations
View
Live Product

REZHIM AI

/ 1.5k+ Users · 5% Conv.

High-load B2C app with 1,500+ organic users. Engineered complex subscription logic, gamification engines, and latency-optimized vision.

Monetization ArchGamificationVision Opt
View

More shipped work

Sales operations

Commercial Proposal Automation

Challenge
Preparing commercial proposals involved client-specific business rules and document formatting.
What I built
Built a generator that applies the client's business rules and produces formatted proposal documents.
Less sales admin time
80%+
Dispatch & field work

Service Operations Automation

Challenge
Dispatchers and field technicians relied on manual data entry to coordinate service work.
What I built
Deployed Telegram-based conversational AI workflows to connect dispatchers and technicians and exchange job details.
Less manual data entry
Fewer entry errors
The Lab
Turn a rough idea into a technical plan you can build from.
Clarify the idea. Answer a few focused questions about the decisions that shape your project.
Explore the architecture. Review a system graph, a suggested stack, and the reasoning behind the choices.
Take the plan with you. Refine the result and download technical documents. This prototype uses model knowledge; live research is not connected.