Shakti S.

Key Skills

Software

Other

OpenCV AI Kit (OAK)

Top Subject Matter

Defense-based machine learning for submarine drag-force prediction (fluid dynamics)

NLP conversational AI with speech-to-text/text-to-speech

Automotive computer vision for driver fatigue detection

Top Data Types

Audio

3D Sensor

Image

Top Task Types

Transcription

Object Detection

Freelancer Overview

Research Intern, DRDO – NSTL: ML research on drag-force prediction for submarines. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Other. Education includes Bachelor of Technology, C.V. Raman Global University and Higher Secondary Certificate (Class XII), Kendriya Vidyalaya No. 2. AI-training focus includes data types such as Computer Code, Programming, and Audio and labeling workflows including Evaluation, Rating, and Transcription.

Labeling Experience

Project: Fatigue Detector App (OpenCV + deep learning)

OpenCV AI Kit (OAK)3D SensorObject Detection

Developed a real-time driver fatigue detection system using OpenCV and deep learning with CNN-based monitoring. Monitored eye blinks and head pose signals to infer driver fatigue conditions under varying lighting. Implemented alert triggers with audio warnings to reduce accident risk based on detection outputs. • Used computer-vision features from video frames for fatigue inference • Trained/evaluated CNN-based detection for eye-blink and pose signals • Calibrated detection behavior to handle different illumination conditions • Connected model outputs to real-time alerting with audio feedback

Not specified

Project: AI Virtual Assistant (NLP + STT + TTS)

OtherAudioTranscription

Built a voice-activated virtual assistant pipeline using NLP along with speech-to-text and text-to-speech components. Implemented intent recognition and context management for multi-turn dialogue to support conversational task execution. Used conversational data processing patterns to enable real-time interaction behavior. • Developed end-to-end audio processing using transcription and spoken responses • Added NLP intent handling and dialogue state/context tracking • Tested multi-turn conversation flows for correctness and relevance • Integrated task-execution behavior with the conversational model

Not specified

Research Intern, DRDO – NSTL: ML research on drag-force prediction for submarines

Performed ML model development for analyzing historical submarine drag-force data derived from fluid-dynamics simulations. Built predictive models to estimate drag force for new submarine designs using ensemble learning methods. Collaborated with defense scientists on preprocessing, feature engineering, and validation of model pipelines for the defense ML research effort. • Worked with structured numeric simulation outputs rather than manual annotations • Applied accuracy-focused evaluation to validate predictive performance • Designed features and validated training data quality through pipeline checks • Supported iterative model improvement with feedback from domain experts

Not specified

Education

Work History