Sakshi Agarwala profile photo 1 on Tal

Sakshi Agarwala

Research | Machine Learning | AI & Intelligent Systems

Bengaluru

4+ years of work experience

Signal Processing Interpretation and REpresentation (SPIRE) Laboratory
Aug 20251 yr 4 mosPresent

Bengaluru, Karnataka, India

Managing and coordinating QA teams for large-scale speech data review, defining quality standards and monitoring consistency across outputs. Designing and maintaining automated data pipelines for speech processing workflows, improving efficiency and reducing manual overhead. Developing and integrating ASR- and VAD-based systems for speech detection, segmentation, and transcription validation. Leading quality analysis and error auditing, using quantitative metrics and region-based review to identify issues and drive improvements. Working on multilingual speech analysis, including language detection and similarity-based evaluations across diverse audio sources.
Signal Processing Interpretation and REpresentation (SPIRE) Laboratory
Oct 202410 mosJul 2025

Bengaluru, Karnataka, India

Deployed and fine-tuned Kosmos, YOLO, and Surya OCR models for document understanding tasks. Built an OCR pipeline with Python, OpenCV, and PyTesseract, automating text extraction and improving data processing efficiency. Evaluated model accuracy using Word Error Rate (WER) and Character Error Rate (CER) via jiwer, optimizing pre-processing for high accuracy. Worked on bounding box detection and explored advanced detection models for structured text extraction. Contributed to LLM pipelines, calculating COMET scores for evaluation and improving text generation quality.
Sep 20232 mosOct 2023

Bengaluru, Karnataka, India

E-Cell Ramaiah
Sep 20222 yrs 11 mosJul 2025

Bengaluru, Karnataka, India

STARDUST_RIT
Feb 20222 yrs 8 mosSep 2024

Bengaluru, Karnataka, India

Sakshi’s AI Stack

antigravity
antigravity
hugging face
hugging face
langchain
langchain
notion
notion

Education

Ramaiah Institute Of Technology

Bachelor of Engineering - BE, Electronics and Instrumentation engineering

Ramaiah Institute Of Technology

20214 yrs2025
Sri Chaitanya College of Education

School, Science

Sri Chaitanya College of Education

20182 yrs2020

Join Sakshi and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Sakshi Agarwala on Tal

Headline

Research | Machine Learning | AI & Intelligent Systems

About

Passionate about research in AI and machine learning, with a focus on building intelligent, data-driven solutions. Constantly exploring deep learning, reinforcement learning, speech technologies, and NLP to solve complex problems and drive innovation. Committed to continuous learning and contributing to impactful projects in the AI and tech space.

Location

Bengaluru

AI stack

antigravity

hugging face

langchain

notion

Work experience

Research Associate at Signal Processing Interpretation and REpresentation (SPIRE) Laboratory, Aug 2025 - Present · 1 yr 4 mos, Bengaluru, Karnataka, India, Managing and coordinating QA teams for large-scale speech data review, defining quality standards and monitoring consistency across outputs. Designing and maintaining automated data pipelines for speech processing workflows, improving efficiency and reducing manual overhead. Developing and integrating ASR- and VAD-based systems for speech detection, segmentation, and transcription validation. Leading quality analysis and error auditing, using quantitative metrics and region-based review to identify issues and drive improvements. Working on multilingual speech analysis, including language detection and similarity-based evaluations across diverse audio sources.

Research Intern at Signal Processing Interpretation and REpresentation (SPIRE) Laboratory, Oct 2024 - Jul 2025 · 10 mos, Bengaluru, Karnataka, India, Deployed and fine-tuned Kosmos, YOLO, and Surya OCR models for document understanding tasks. Built an OCR pipeline with Python, OpenCV, and PyTesseract, automating text extraction and improving data processing efficiency. Evaluated model accuracy using Word Error Rate (WER) and Character Error Rate (CER) via jiwer, optimizing pre-processing for high accuracy. Worked on bounding box detection and explored advanced detection models for structured text extraction. Contributed to LLM pipelines, calculating COMET scores for evaluation and improving text generation quality.

Mobile Application Developer at Polism Solutions Private Limited, Sep 2023 - Oct 2023 · 2 mos, Bengaluru, Karnataka, India

Executive Committee Member at E-Cell Ramaiah, Sep 2022 - Jul 2025 · 2 yrs 11 mos, Bengaluru, Karnataka, India

Executive Committee Member at STARDUST_RIT, Feb 2022 - Sep 2024 · 2 yrs 8 mos, Bengaluru, Karnataka, India

Education

Bachelor of Engineering - BE, Electronics and Instrumentation engineering, Ramaiah Institute Of Technology

School, Science, Sri Chaitanya College of Education

Public links

github: https://github.com/sakshiagarwala0206

Sakshi Agarwala - Research Associate in Bengaluru | Tal