Anjali T.

Key Skills

Software

Other

Top Subject Matter

Diabetes prediction and explainable

fair clinical AI

Meeting audio to diarized transcription and minutes generation

Top Data Types

Audio

Text

Image

Top Task Types

Diagnosis

Transcription

Freelancer Overview

DiabetesXAI-Trust: Explainable AI for Diabetes Prediction. Core strengths include Other. Education includes Master of Science, Himachal Pradesh University (2026) and Bachelor of Science, Himachal Pradesh University (2023). AI-training focus includes data types such as Medical, DICOM, and Audio and labeling workflows including Diagnosis and Transcription.

Labeling Experience

Speech-to-MoM: Audio Intelligence Application

OtherAudioTranscription

Created an end-to-end audio intelligence workflow that transforms raw meeting audio into structured meeting minutes and summaries. Used transcription and speaker diarization to generate labeled, time-anchored conversational content suitable for downstream text generation. • Used PyAnnote.audio for speaker diarization to label who spoke when. • Used WhisperX to transcribe audio into text. • Integrated LangChain with an LLM backend to generate role-attributed meeting summaries from transcripts. • Built storage and serving components to manage processed audio metadata and outputs.

2026 - 2026

DiabetesXAI-Trust: Explainable AI for Diabetes Prediction

OtherDiagnosis

Built an explainable diabetes prediction pipeline using supervised learning outputs to support clinical decision support. Focused on producing trustable model explanations and fairness-aware reliability metrics for patient groups. • Applied SHAP and LIME to generate feature-level explanations for model decisions. • Implemented conformal prediction to quantify uncertainty and support reliability. • Designed a composite trust score combining performance and uncertainty signals. • Evaluated fairness using demographic parity and equalized odds across demographics.

2026 - 2026

Education