Mrinaal.M Harihar

About

Software engineering student graduating in 2027 with 8 months of remote startup experience shipping production systems end to end. Builds full-stack services in Python and TypeScript, and develops ML and LLM pipelines with a focus on maintainable, testable, and fault-tolerant systems.

Experience

• Software Engineering Intern
  ThoughtWire.AI
  Dec 2025 – Jul 2026

Education

• B.E.
  RV Institute of Technology and Management
  Information Science and Engineering · 2027

Skills

• SQL
• Computer Networks
• Express.js
• FastAPI
• Python
• CSS3
• Java
• JavaScript
• Node.js
• Tailwind CSS
• NumPy
• Pandas
• scikit-learn
• React
• PyTorch
• TensorFlow
• Responsive Design
• TypeScript
• Operating Systems
• Next.js
• OpenAI API
• OAuth
• JWT
• Microservices
• Scrum
• API Integration Understanding
• CI/CD
• Secure Coding
• HTML5
• Design Systems
• SDLC
• Design Patterns
• LangGraph
• Data Structures and Algorithms
• MVC Architecture
• Pydantic
• C/C++
• SQLAlchemy
• Oop
• Google Adk
• Advanced DBMS
• uvicorn

Projects

• AETHER-CRYPTO – 128-bit SPN Block Cipher
  Python, Algorithm Design

  Implemented a block cipher from first principles, including S-box construction, permutation layers, and key scheduling.

• KuChiKu – AI Interview and Evaluation Platform
  React, Node.js, Express.js, Supabase

  Built a full-stack interview and evaluation platform with LLM-generated questions, real-time scoring, authentication, and modular services.

• Contract Risk Analyzer – Agent Training Environment
  PyTorch, Docker, FastAPI

  Built an OpenEnv-compatible environment with versioned action/observation APIs and tiered F1-based evaluation across 13 clause risk types; containerized and deployed to Hugging Face Spaces.

• LugWatch – Agentic Competitive Intelligence
  LangGraph, Ollama, FastAPI, SQLAlchemy, pytest

  Architected a three-stage LangGraph agent state machine for external data collection and cross-brand recommendations, with local Mistral 7B serving, pytest coverage, and offline fallback.

• Kavach AI – Production ML Platform
  FastAPI, React 19, PyTorch, ONNX, GraphSAGE

  Built a four-module anti-fraud platform with reproducible training scripts, ONNX CPU inference, GraphSAGE-led money-mule detection, and a health endpoint that reports true loaded-model state.

Courses & certifications

• Low Code / No Code Workshop · ClassVerse · 2025
• IBM AI and Cloud Computing · IBM · 2024
• Tableau Data Visualization · 2024

🎯 Hobbies & interests

• GitHub repositories
• Smart Draft
• Vektor
• Gmail Automation Agent