Project Overview
PRISM is an end-to-end AI recruitment platform designed to solve the operational complexity HR teams face when managing large-scale hiring. Traditional hiring processes rely heavily on manual coordination across resumes, interviews, and feedback collection. PRISM replaces this fragmented system with a unified AI-driven workflow that automates resume screening, interview scheduling, AI-led screening rounds, candidate monitoring, and structured evaluation.
System Architecture
PRISM combines workflow automation with machine learning and computer vision to create an intelligent hiring pipeline. The system integrates resume analysis, automated scheduling, AI interviews, and behavioral monitoring into one platform.
AI Resume Screening Engine
Uses NLP to analyze resumes against job descriptions and generate a JD Fit Score that ranks candidates automatically.
Smart Interview Scheduling
Connects to interviewer calendars and automatically finds optimal interview slots, eliminating manual coordination.
AI Interview Engine
Conducts conversational AI screening interviews with candidates, generating transcripts and response scoring.
Computer Vision Monitoring System
Uses YOLOv8 and MediaPipe to track candidate behavior during interviews, monitoring face presence, gaze direction, and anomalies.
Structured Evaluation System
Collects standardized feedback from interviewers with scoring frameworks tied directly to job requirements.
Implementation Details
Code Example
# YOLO Person Detection
from ultralytics import YOLO
model = YOLO("yolov8n.pt")
results = model(frame, classes=[0])
person_count = len(results[0].boxes)
# MediaPipe Face Mesh
import mediapipe as mp
face_mesh = mp.solutions.face_mesh.FaceMesh(
refine_landmarks=True,
max_num_faces=10
)
# Eye Aspect Ratio (Blink Detection)
ear = (vertical1 + vertical2) / (2 * horizontal)
if ear < 0.19:
eye_status = "EYES CLOSED"
else:
eye_status = "EYES OPEN"
# Head Pose Estimation
success, rotation_vec, translation_vec = cv2.solvePnP(
model_points,
image_points,
camera_matrix,
dist_coeffs
)Agent Memory
PRISM smooths gaze detection using a rolling buffer of eye ratios to avoid false movements caused by small eye jitter or camera noise.
Workflow
PRISM transforms the recruitment pipeline into an automated AI workflow where candidates move through application, AI screening, interview scheduling, evaluation, and offer generation without manual coordination.
Results & Impact
"PRISM removed the operational chaos from our hiring process. What used to take hours of coordination now happens automatically."
85% Reduction in Manual HR Work
Automated resume screening, scheduling, and feedback collection eliminated repetitive administrative tasks.
Faster Candidate Decisions
AI screening and instant scheduling reduced the time between application and decision significantly.
Improved Hiring Quality
Structured feedback and AI candidate scoring enabled more consistent and data-driven hiring decisions.
Interview Integrity Monitoring
Computer vision analysis detects anomalies such as multiple people in frame or excessive off-screen gaze.
About the Author
Contributors
Ready to Build Your AI Solution?
Get a free consultation and see how we can help transform your business.