RAM RITHEESH ramritheesh

👨💻 About Me

class RamRitheesh:
    def __init__(self):
        self.role      = "AI/ML Engineer (in the making)"
        self.studying  = "B.Tech, Robotics and AI @ BIT"
        self.cgpa      = 8.63
        self.location  = "Bengaluru, India"
        self.loves     = ["ML pipelines", "LLM apps",
                          "Explainable AI", "Clean APIs"]
        self.looking_for = "AI/ML internship"

    def mission(self):
        return "Turn messy data into models people can trust."

🔭 Building: LLM, RAG and fraud-detection systems

🧪 Obsessed with: model evaluation and explainability

⚙️ Strength: end-to-end, from SQL and models to APIs and UI

🎯 Goal: ship production-ready AI at a product team

🧰 Tech Arsenal

🚀 Featured Projects

🛡️ RazorGuard: fraud detection that explains itself

Rule-based fraud limits miss behavioral patterns. RazorGuard scores each transaction against the customer's own history.

🎯 ROC-AUC	⚖️ Threshold	🔁 Recall	⚡ Latency
0.9055	0.18 (cost-sensitive)	61.9%	< 200 ms

flowchart LR
    A["Transaction"] --> B["Feature engineering<br/>amount ratio, merchant velocity"]
    B --> C["XGBoost classifier"]
    C --> D["SHAP explanation"]
    C --> E{"Risk score 0-100"}
    E -->|low| F["ALLOW"]
    E -->|medium| G["REVIEW"]
    E -->|high| H["BLOCK"]
    D --> I["React dashboard"]
    E --> I

🔍 More project highlights (click to expand)

🧭 Career.OS: AI career assistant Ranks jobs with a hybrid score: 40% keyword + 60% semantic similarity (Hugging Face embeddings, cosine similarity). A RAG pipeline powered by Gemini gives personalized career chat, and the "Explain My Match" feature shows skill gaps and strengths. JWT auth with per-user data isolation. React Node.js Express MongoDB Atlas Gemini API RAG

🚨 ScamShield AI: digital safety agent Paste a suspicious SMS, email or screenshot and get a 0-100 risk score, a classification and clear next steps. Async FastAPI backend with structured Gemini prompts and multimodal screenshot analysis. API keys stay server-side. FastAPI Gemini API React Render Vercel

📄 InsurAI Classifier: claims triage Hybrid NLP and generative AI pipeline that classifies unstructured insurance claim documents and extracts policy entities and triage tags. TypeScript React NLP Gemini API

💼 Experience

Project Intern · KodNest Technologies | Feb 2026 to May 2026

⚙️ Built FastAPI microservice components with Pydantic contracts between services
🧹 Automated ingestion validation that filtered malformed inputs with 95% precision
🧪 Wrote Pytest unit and regression suites that caught 25+ integration bugs before staging
📊 Built Power BI and Excel reports for team leads; documented runbooks and test logs on Git/GitHub

🎓 Education and certifications

B.Tech, Robotics and Artificial Intelligence, Bangalore Institute of Technology (2022 to 2026), CGPA 8.63/10
AI/ML Specialization, Infosys
Introduction to Machine Learning, NPTEL
Full-Stack Web Development, Internshala
AI Foundation Course, Jio Institute

📊 GitHub Analytics

🤝 Let's Connect

Hiring for AI/ML, or building something with LLMs? Let's talk.

"Models are easy to build. Models people can trust are the real work."

Navigation Menu