Skip to content
View amoshnin's full-sized avatar
🀩
🀩

Highlights

  • Pro

Organizations

@Entrepreneurship-Bank @GlowCar-Marketplace @NUS-MedVision @Scootify-App @University-Modules @Work-Applications

Block or report amoshnin

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
amoshnin/README.md

Banner


My GitHub Organizations

πŸ“¦ Collections of Repositories

Repositories I maintain or organize - personal incubators, university modules and work-related projects.

  • Artem's Entrepreneurship-Bank (org): My personal incubator for early-stage projects - experimenting & prototyping before ideas evolve further.
  • Artem's University Modules (org): Where I keep code for my university courses β€” a compilation of projects & experiments from my studies.
  • Artem's Work Applications (org): Collection of real-world software projects developed for clients and companies across different domains.

πŸ› οΈ Engineering Challenges

Complex systems, scalable architectures and AI pipelines I've architected to solve hard technical problems.

  • Fluxo (website): A fully operational, production-ready SaaS platform engineered for compliance-sensitive AI workflow orchestration. It empowers agencies and operations teams to build complex, composable node graphs using a natural-language to visual builder, seamlessly connecting AI models, logic routing, and external integrations like Postgres and Slack. To solve the "black box" problem of unchecked AI, Fluxo strictly enforces Human-in-the-Loop (HITL) governance through policy-based routing that can auto-approve safe actions, block errors, or escalate risky outputs to a centralized Review Inbox. The platform boasts advanced features including AI prompt learning that automatically suggests system prompt improvements based on human edits, live execution analytics & white-labeled client portals for secure, external sign-offs via custom domains. Backed by enterprise-grade controls like scoped API keys, signed webhooks, role-based access & multi-tier subscription billing, Fluxo ensures teams can scale autonomous workflows rapidly without ever losing auditability or human approval control.

  • LedgerX (website | repository): A robust financial integrity engine built to manage high-concurrency transaction environments. Architected with strict ACID-compliant database locking and immutable audit trails to guarantee absolute mathematical accuracy and prevent race conditions at scale.

  • DocuMind (website | repository): A document intelligence platform featuring a custom Hybrid RAG pipeline (Vector + BM25) and dynamic LLM routing. It leverages a secure, stateless Bring Your Own Key architecture to parse private enterprise PDFs without leaking data to corporate training sets.

πŸ’‘ Entrepreneurial Projects

Startups and serious projects I built from scratch or developed through my incubator.

  • GlowCar (org): A second-hand car marketplace connecting buyers and sellers β€” grew out of my incubator to become a significant startup project.
  • Scootify (org): Sustainable urban mobility platform β€” green, efficient & connected through shared electric scooters.

πŸŽ“ Other Projects

Other projects I contributed to as part of my studies or extracurricular work.

  • MedVision+ (org): Full-Stack diagnostic assistant built as my NUS final year project - supporting multi-modal medical scans with deep learning & LLMs.

banner bottom

Pinned Loading

  1. Business.Project-LedgerX Business.Project-LedgerX Public

    A high-concurrency financial integrity engine featuring strict ACID-compliant database locking and immutable audit trails.

    TypeScript 1

  2. Business.Project-DocuMind Business.Project-DocuMind Public

    A vendor-agnostic document intelligence platform with a Hybrid RAG pipeline and a stateless BYOK architecture for secure PDF analysis.

    TypeScript 1

  3. ML-Human.Activity.Recognition ML-Human.Activity.Recognition Public

    My machine learning pet project on Human Activity Recognition Using Multiclass Classification with fitness data from a smartphone tracker

    Jupyter Notebook 1

  4. ML-Topic.Modelling.to.Group.Customer.Reviews ML-Topic.Modelling.to.Group.Customer.Reviews Public

    My machine learning pet project on topic modelling using K-Means clustering to group customer reviews

    Jupyter Notebook 1 1

  5. ML-Credit.Card.Fraud.Detection.Anomaly.Detection ML-Credit.Card.Fraud.Detection.Anomaly.Detection Public

    My machine learning pet project on Credit Card Fraud Detection as an Anomaly Detection Problem with Autoencoders

    Jupyter Notebook 1

  6. ML-Credit.Card.Fraud.Detection.Classification ML-Credit.Card.Fraud.Detection.Classification Public

    My machine learning pet project on Credit Card Fraud Detection as a Classification Problem

    Jupyter Notebook 1 1