Looking for a job

Debanjan Chakraborty

@debanjanchakraborty1

Message

AI Trainer and LLM Evaluation Specialist across RWS and Soul AI, reviewing model outputs and supporting RLHF training.

India

Message

Across freelance projects with RWS and Soul AI, I’ve evaluated AI and speech outputs, including Bengali-language work, and supported LLM training through prompt review and response evaluation.

At AKOVA CONSULTANCY PRIVATE LIMITED, I built AI solutions for FASTag and intelligent transport systems, including models for sales-trend prediction and distributor/agent performance analysis.

I also developed computer-vision systems for number-plate recognition and lane detection, and integrated AI models with production applications using Docker, Firebase, and AWS. At Dezzex Technologies, I fine-tuned a YOLOv8 segmentation model for road-asset detection in satellite imagery.

Experience

Work history, roles, and key accomplishments

Current

Freelance AI Trainer & Evaluator

Current

Freelance

Jan 2023 - Present (3 years 9 months)

Provided AI training, LLM evaluation, RLHF, data annotation, transcription, and video annotation services for various freelance projects.

LLM Evaluation RLHF AI Training Data Annotation Transcription Video Annotation Prompt Generation Quality Assessment

AI/ML Engineer

AKOVA CONSULTANCY PRIVATE LIMITED

Jan 2025 - Dec 2025 (11 months)

Built end-to-end AI solutions for FASTag and intelligent transport systems. Developed ML models including XGBoost and LSTM for sales-trend prediction and performance analysis, and designed YOLOv8/OpenCV computer-vision systems for number-plate recognition and lane detection.

XGBoost LSTM YOLOv8 OpenCV Computer Vision Docker Firebase AWS MLOps

Machine Learning Intern

Dezzex Technologies

Sep 2024 - Nov 2024 (2 months)

Developed and fine-tuned a YOLOv8 segmentation model for road-asset detection in satellite imagery. Used Roboflow for annotation and preprocessing, achieving Box mAP50 0.604 and Mask mAP50 0.311.

YOLOv8 Segmentation Roboflow Satellite Imagery Annotations Pre processing

Data Science Intern

iNeuron Intelligence Pvt Ltd

Dec 2023 - Jan 2024 (1 month)

Processed, cleaned, and modeled data using Python, Pandas, and NumPy. Built credit-risk models using Logistic Regression, Random Forest, and XGBoost and deployed a monitoring/retraining pipeline.

Python Pandas NumPy Logistic Regression Random Forest XGBoost Credit Risk Modeling Monitoring Retraining Pipelines

AI Data Annotation

Innodata

Performed AI data annotation and quality-review tasks using structured project workflows and guidelines.

Data Annotation Quality Review

AI / Speech Evaluation

RWS

Performed AI and speech evaluation, transcription, and quality-review tasks, including Bengali-language work. Reviewed model outputs against task-specific criteria and project guidelines.

AI Evaluation Speech Evaluation Transcription Quality Review Bengali

LLM Training / RLHF

Soul AI

Supported LLM training by creating and reviewing complex prompts and evaluating AI-generated responses. Conducted research and provided quality-focused feedback to improve model accuracy.

LLM Training RLHF Prompt Generation Prompt Review AI Response Evaluation Research

Video Annotation

Highbrow Technology Inc.

Performed video annotation and structured AI-training tasks. Produced voiceover video walkthroughs covering Microsoft 365, Google Workspace, AWS, Firebase, and VS Code.

Video Annotation AI Training Voice Over Microsoft 365 Google Workspace AWS Firebase VSCode