AI/ML Engineer with hands-on experience designing LLM-powered multi-tool agents, RAG pipelines, and end-to-end workflow automation systems. Engineered & deployed 3 interactive AI applications — achieving a 70% reduction in manual review time and 85% improvement in response accuracy. Anthropic-certified · Building with the Claude API. Actively seeking AI/ML Engineer Internship roles.
I'm a Computer Science & Engineering student with deep, hands-on focus on AI/ML engineering — not just theory, but real production systems. I design LLM-powered multi-tool agents, RAG pipelines, and end-to-end workflow automation that solve genuine problems at scale.
My work lives at the intersection of LLM orchestration, healthcare AI, and applied ML. I've engineered and deployed 3 interactive end-to-end AI applications — Clinexa AI, MediBot, and MediFlow AI — achieving a 70% reduction in manual review time and 85% improvement in domain-scoped response accuracy. I also build classical ML pipelines spanning fintech risk and real estate price prediction.
I'm certified in Building with the Claude API (Anthropic), Prompt Engineering (IBM), and AI Foundations (Simplilearn). I'm always pushing the frontier of what's possible with LangChain, LlamaIndex, Gemini API, n8n, and modern ML tooling.
From prompt engineering to production pipelines — the full stack I use to build and ship AI applications.
All AI projects are production-deployed with live demos. ML projects include full end-to-end pipelines with deployed dashboards.
Full-stack AI HMS integrating Gemini API + LangChain across 6+ hospital workflows: patient report analysis, appointment priority detection, AI doctor notes, medical chatbot — all in one deployable system. Firebase admin dashboard with role-based access control.
Multi-tool LangChain agent (ZERO_SHOT_REACT_DESCRIPTION) with Gemini API for clinical query resolution, appointment lookups, and patient summarization. Autonomously selects tools based on user intent — true agentic behavior with ConversationBufferMemory for stateful multi-turn conversations.
End-to-end patient intake automation pipeline: form submission → Gemini AI-powered clinical summary → structured notification dispatch. Eliminates manual data routing across hospital systems. Extensible modular architecture with multi-step webhook-triggered workflows.
Enterprise-grade ML platform predicting microfinance loan repayment behavior using telecom customer data. Identifies potential defaulters and improves credit decision-making through AI-driven analytics. Full end-to-end pipeline with batch prediction, CSV upload, and downloadable reports.
Regression ML model with feature engineering across 15+ variables achieving 92% prediction accuracy. Automated preprocessing pipeline cut analysis setup time by 30%. Interactive Streamlit dashboard with Gemini API layer for natural language explanations of individual predictions.
End-to-end ML pipeline for predicting Australian housing prices — built for Surprise Housing's market entry strategy. 1,460 training records, 81 features, advanced regression modeling, EDA, feature engineering, and investment-focused business insight extraction.
Snapshot of my technical background, certifications, and key achievements.
Open to AI internships, full-time roles, freelance projects, and open-source collaborations. I reply within 24 hours.