All ProjectsMahesh.
01AI Talent & Career Platform

AI-Based Resume Evaluator & Interviewer

Accelerates candidate hiring with automated ATS scoring and real-time interactive mock interviews — bridging the gap between static resumes and live evaluation. AI-powered platform that evaluates resumes against job descriptions, provides ATS optimization suggestions, and conducts adaptive mock interviews with real-time scoring and feedback.

Next.jsTypeScriptLangGraphGenerative AITailwind CSS
View Code

AI-Based Resume Evaluator & Interviewer

The AI Resume Evaluator and Interviewer is a comprehensive career acceleration platform. By leveraging generative AI (LLMs) and advanced state machines, it bridges the gap between static resume scoring and dynamic, real-time interview coaching.

Architecture & Data Flow

At its core, the platform uses a React (Next.js) frontend and a highly concurrent backend orchestrated by LangGraph. The workflow operates in three distinct phases:

1. Document Ingestion & Parsing: The system accepts PDF/Docx files, strips non-essential formatting, and extracts raw text using Optical Character Recognition (OCR) fallback mechanisms for image-heavy resumes.

2. ATS Scoring & Gap Analysis: The extracted resume text is compared against a user-provided job description. An LLM agent (equipped with strict structured output schemas) scores the resume on keyword matching, semantic relevance, and structural clarity.

3. Adaptive Mock Interviews: The crown jewel of the platform is the interactive interview loop. Based on the identified gaps in the resume, the system dynamically generates behavioral and technical questions.

State Management with LangGraph

Handling multi-turn conversational agents requires robust state management to prevent runaway context windows. We utilize LangGraph to enforce deterministic state transitions:

  • AgentState: Maintains the history of questions asked, current user responses, and cumulative score.
  • Evaluation Node: Analyzes the user's latest response in real-time, assigning a score out of 10 and determining if a follow-up clarification is needed.

Security & Scalability

  • Data Privacy: Resume data is processed in memory and immediately discarded. PII (Personally Identifiable Information) masking is applied before sending payloads to external LLM providers.
  • Rate Limiting: To prevent API abuse, rigorous IP-based and token-based rate limiting is implemented via Upstash Redis.
Explore More Work
02 · Real-Time Cross-Device ClipboardCase Study

ClipYard

Real-time cross-device clipboard app enabling account-free text sync and P2P file transfer via WebRTC, secured with AES-256-GCM encryption and JWT auth.

03 · Visual Python IDE & AST EngineCase Study

LogicPup

Full-stack visual Python IDE with a custom AST engine converting flowcharts to runnable Python, an 8-level gamified curriculum, and OAuth backed by PostgreSQL.

AI-Based Resume Evaluator & Interviewer — Mahesh | Mahesh Chopade Portfolio