AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI Engineer with 1+ years in Healthcare AI platforms & Full-Stack Development
AI Engineer with production experience building healthcare AI platforms at Zauto AI. Designed and shipped a multi-channel AI Healthcare Agent (Voice, WhatsApp, Web) automating appointment booking and real-time patient record retrieval using custom RAG pipelines and LLM function calling across 20+ hospitals. Led on-site deployments integrating real-time doctor-patient conversation AI into live EMR systems. Independently architected an enterprise-grade codebase intelligence system using advanced retrieval techniques including Multi-Stage Retrieval, BGE-Reranker v2, AST-based chunking, and semantic caching – achieving 95%+ retrieval precision while reducing token costs by 30%. Hands-on with MCP (Model Context Protocol) for tool-augmented agentic workflows, LangChain/LangGraph orchestration, and production AI infrastructure built with Python, FastAPI, and NestJS/Node.js. Strong understanding of LLM orchestration, multimodal AI systems (TTS, STT, Embeddings), vector database architecture, and scalable AI application design.
Karpagam Academy of Higher Education
B.Tech · Artificial Intelligence & Data Science
August 1, 2020 – June 30, 2024
Zauto AI
AI & Full-Stack Engineer
January 1, 2025 – Present
Salem, Tamil Nadu, India
Alhena - Healthcare AI Automation Platform
June 24, 2026 – Present
Engineered end-to-end Voice-EMR automation retrieving lab reports, prescriptions, and visit history via custom RAG pipelines on Milvus — handling 500+ patient queries daily across Telephony and WhatsApp with sub-second retrieval. Implemented Tiktoken-based dynamic chunking with metadata filters for patient-record queries — achieving ~40% reduction in clinical hallucination rate vs fixed-size chunking baseline. Integrated real-time EMR data access via LLM Function Calling — enabling fully automated appointment booking and dynamic service updates, eliminating manual staff intervention for routine queries.
LLM Gateway - Centralised AI Model Management Platform
June 24, 2026 – Present
Architected a centralised LLM gateway with unified multi-provider routing across Text, TTS, STT, and Embedding modalities — supporting OpenAI, Azure, ElevenLabs, Sarvam, Deepgram, and others with real-time streaming. Implemented capability-aware model routing with NER-based guardrails, per-org API key isolation, and token & rate limiting - cutting average cost-per-query by ~20% through intelligent provider arbitrage.
AI Prompt Manager - Agent-Oriented Prompt Orchestration Platform
June 24, 2026 – Present
Engineered an agent-based prompt orchestration system supporting multi-LLM routing, semantic version control, and runtime configuration hot-swaps — managing 50+ active prompt templates across 4 AI agents in production. Implemented a canary-style rollout mechanism for gradual traffic shifting between prompt versions — reducing mean rollback time from hours to under 5 minutes and eliminating prompt regression incidents in production.
Meta WhatsApp Integration Platform - In-House Communication Infrastructure
June 24, 2026 – Present
Single-handedly delivered an in-house Meta WhatsApp Business platform replacing third-party SaaS providers — with full messaging, template, and account management capability. Built event-driven webhook infrastructure processing 2,000+ events/min at sub-200ms delivery latency — eliminating recurring external vendor costs.
Enterprise Repo-RAG — Agentic Codebase Intelligence System
June 24, 2026 – Present
Architected an Agentic RAG system using LangChain AgentExecutor with Tavily Web Search and custom code-tools for multi-source intelligence across KT Docs, Source Code, and Live Documentation — built independently to deepen advanced RAG and agentic workflow knowledge. Built a Multi-Stage Retrieval pipeline (Parent Document Retrieval + Multi-Vector) refined by BGE-Reranker v2 — achieving 95%+ retrieval precision on codebase queries. Implemented AST-based language-aware chunking, Session Namespacing for multi-tenant retrieval, and pgvector-backed Semantic Caching — reducing token costs by 30%. Deployed open-source model inference (Llama 3.1, Qwen) via Hugging Face and Groq achieving sub-300ms latency; built observability stack with Prometheus and Grafana monitoring P99 latency and RAGAS Faithfulness/Relevancy scores in real time.
Foundation: Introduction to LangChain - Python
LangChain Academy
May 1, 2026 – Present
Quickstart: LangChain Essentials - Python
LangChain Academy
May 1, 2026 – Present
Foundation: Introduction to LangGraph - Python
LangChain Academy
May 1, 2026 – Present
Cultural Fit Analysis
The candidate's project portfolio showcases a strong alignment with the target role of an AI Engineer, particularly in advanced RAG, agentic systems, and LLM orchestration. The mix of professional and self-initiated projects demonstrates initiative, continuous learning, and a passion for AI. Their experience in healthcare AI indicates an ability to apply technical skills to complex, high-impact domains. The breadth of technologies and frameworks used suggests adaptability and a willingness to explore different solutions, which is a good indicator of cultural fit within a dynamic engineering environment.
Soft Skills & Operational Fit
The candidate demonstrates strong problem-solving skills through their project descriptions, tackling complex issues like clinical hallucination reduction, cost optimization, and real-time data integration. Their experience with canary-style rollouts and observability indicates a focus on robust, production-ready systems and operational excellence. The self-initiated 'Enterprise Repo-RAG' project highlights a proactive learning attitude and deep technical curiosity, which are valuable for cultural fit in an innovative AI team.