Shashikant Kumar B.

Key Skills

Software

Other

Roboflow

Data Annotation Tech

Top Subject Matter

Financial document information extraction for unstructured data to structured fields

Real-time IoT water quality monitoring and contamination risk forecasting

Top Data Types

Document

Image

Computer Code Programming

Top Task Types

Function Calling

Data Collection

Fine-tuning

Text Generation

Computer Programming/Coding

Freelancer Overview

Summer Research Intern Mentor. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Bachelor of Technology, University of Delhi (2027) and Bachelor of Science, IIT Madras (2028). AI-training focus includes data types such as Document and labeling workflows including Function Calling and Data Collection.

Labeling Experience

Data Analyst Intern (Remote) — Soul AI (April 2025 – Present)

OtherDocumentFunction Calling

Built an automated pipeline to extract financial information from unstructured documents using LLaMA 3 and Gemini APIs. Evaluated and iterated model outputs to improve extraction quality and reduce human effort. Focused on prompt/function-oriented workflows to drive consistent structured results. • Extracted relevant fields from unstructured inputs • Worked with LLM APIs for document understanding • Performed iterative accuracy improvements to reduce manual processing time • Supported automated downstream financial data structuring

2025 - Present

Summer Research Intern — DAViSE Lab, NIT Delhi (June 2025 – Aug 2025)

OtherDocumentData Collection

Designed an end-to-end IoT edge-to-cloud processing approach for real-time water quality monitoring across multiple sensor types. Developed deep learning forecasting models (LSTM and SARIMA) to predict contamination risk using collected sensor data. Implemented an edge preprocessing pipeline to reduce redundant transmissions and support timely alerting. • Architected Edge-Fog-Cloud workflow for real-time monitoring • Built LSTM/STARIMA-based contamination forecasting models • Implemented edge preprocessing to reduce redundant cloud data by 35% • Integrated ML outputs into a React-based GIS dashboard for real-time alerts

2025 - 2025

Education

Work History