1Git2Clone
AI Integration Developer | APIs, Data Pipelines & Internal Tools
I turn operational workflows into production features that are useful, auditable, and maintainable.
I build API backends, AI-backed dashboards, document processing tools, data ingestion flows, automation scripts, and developer tooling. Recent production work spans OpenAI, Gemini, RAG/code graph systems, OCR, computer vision, and AWS services like Lambda, S3, IAM, VPC, CloudWatch, and CloudFormation.
I use Go professionally, prefer Rust for systems work, and build interfaces with Svelte or React. I have contributed to rust-random/rand, the standard Rust random number generation library used across the ecosystem.
Available remotely for AI integrations, API creation, data pipelines, internal tools, and backend infrastructure.
If you need an AI feature, API, dashboard, or data pipeline turned into something that can run outside a demo, get in touch. ivan@hu-tao.org
Async-friendly. I work from rough requirements to shipped tool. Remote-first, available for new contracts.
Production AWS Lambda/S3 · rust-random/rand contributor · 20x benchmark improvement · Sub-100ms Discord bot · PCI-adjacent fintech
API Creation
Go services and AWS Lambda pipelines that handle async workflows, data ingestion, and backend glue for production systems.
AI Integration
LLM features, RAG systems, document processing, and AI-backed automation that remain useful outside the demo.
Internal Dashboards & Tooling
Svelte/React admin interfaces, observability dashboards, and operational tools that teams actually use.
Data Pipelines
ETL flows, streaming consumers, and analytics infrastructure that scales from batch to event-driven workloads.
Work Beyond Home (localhost)
Things that actually shipped.
Fintech Transaction Pipeline (Confidential)
A production fintech data pipeline for safely processing and analyzing transaction flows.
Tech Stack:
- Compute: AWS Lambda
- Storage: Amazon S3
- Languages: Go
Problem
Process sensitive transaction data for analytics while guaranteeing correctness under failure conditions.
Solution
Built a greenfield AWS Lambda with NDJSON streaming to redact transaction data under memory constraints, feeding a consumer flow that analyzes, deletes, and preserves insights.
Outcome
Enabled automated failure-rate analytics on production flows with 20x benchmark improvements while maintaining correctness, auditability, and data-handling compliance.
AI Admin Tools (Confidential)
Internal AI tooling for operational analytics, policy scanning, and document forensics.
Tech Stack:
Problem
An internal admin system needed AI tools for analyzing failures, scanning policies, and inspecting documents without regex-only brittle workflows.
Solution
Built AI-backed features for the Go dashboard: failure-rate insights, an AI policy scanner that handles SPAs, and document forensics using ExifTool, Tesseract, OpenCV, and Gemini computer vision.
Outcome
Replaced fragile review workflows with auditable, reusable tools that help teams inspect operational issues, policy changes, and suspicious metadata faster inside production systems.
Serenity Discord Bot
Asynchronous Discord bots with reactive database integrations and sub-100ms response times.
Tech Stack:
- Backend Libraries: Serenity, Tokio
- Databases: PostgreSQL
- Languages: Rust
Problem
Standard Discord bot frameworks trade off response time or require verbose async handling, making it hard to build responsive bots without boilerplate.
Solution
Leveraged Serenity's event-driven architecture with Tokio for non-blocking I/O, PostgreSQL for state, and optimized task queuing to handle rapid user interactions.
Outcome
Deployed bots achieving sub-100ms response times in production, handling thousands of concurrent users with minimal latency.
Got A Messy Idea Worth Shipping?
Send me the rough version. I'll help turn it into something real: ivan@hu-tao.org.