Suganya P

GenAI Systems Engineer | AWS Bedrock | LLM Systems | RAG & Orchestration

Jan 2025 - Present • 1 yr 5 mos

Agentic-RAG-Resume-Intelligence Built a governance-aware AI resume intelligence platform using LangGraph, Flask, FAISS, and Sentence Transformers to move beyond traditional keyword-based ATS systems through contextual semantic retrieval and deterministic evaluation workflows.  Developed modular LangGraph orchestration pipelines for parsing, extraction, retrieval, analysis, and evaluationDeveloped PDF parsing and competency extraction pipelines.  Replaced keyword ATS matching with semantic competency-based retrieval using Sentence Transformers and FAISS.  Implemented weighted semantic scoring for realistic recruiter-style candidate evaluation  Added governance controls including PII masking, workflow validation, and safety filtering  Integrated evaluation frameworks like LangSmith, RAGAS, and DeepEval for retrieval quality and scoring validation  Designed explainable recruiter-facing outputs including strong matches, missing skills, retrieval precision, and score consistency Technologies : Python, Flask, LangGraph, LangChain, FAISS, Sentence Transformers, LangSmith, RAGAS, DeepEval AI Companion System for Guided Conversations – (Bro-Buddy) Production-oriented GenAI workflow built using AWS Bedrock and serverless architecture with focus on reliability, observability, and controlled execution.  Designed and deployed a serverless GenAI system using AWS Bedrock, Lambda, and API Gateway  Built deterministic routing workflows reducing unnecessary LLM calls by 30–40% and improving latency  Implemented multi-layer safety pipeline with PII masking, prompt injection filtering, validation, and fallback handling  Developed orchestration flow integrating routing, memory handling, validation, and response control  Built structured observability for latency tracking, routing decisions, token monitoring, and debugging  Achieved 92% routing accuracy with controlled AI execution and governance-focused workflow design Technologies : AWS B

AI Systems Reliability Engineer

May 2023 - Mar 2024 • 10 mos

Monitored AI system performance using CloudWatch and structured logging  Supported backend debugging and routing issue analysis for workflow stability  Improved operational reliability through structured validation and issue tracking  Collaborated with backend and AI teams to analyze response inconsistencies and deployment issues