NLP
Flunkey
Sentiment Analysis of Customer Reviews
Built comprehensive sentiment analysis system using BERT fine-tuning and traditional ML approaches. Implemented data preprocessing with NLTK, feature extraction using TF-IDF and Word2Vec, and SQLite database integration. Achieved high accuracy sentiment scoring with multi-approach validation for customer feedback analysis.
8 months
Project Duration
High accuracy sentiment scoring
Key Achievement
6+
Technologies Used
The Challenge
Insight Fusion Analytics needed an autonomous vehicle steering prediction system capable of real-time decision-making with high precision for self-driving applications.
The Solution
Created an LSTM-CNN hybrid architecture optimized for NVIDIA DRIVE PX platform, integrated with YOLO for object detection and TensorRT for inference optimization across multi-camera inputs.
Results & Impact
- Achieved <1.5° mean absolute error in steering prediction
- Optimized inference latency by 40%
- Enabled 60 FPS real-time decision-making
- Successfully integrated multi-camera sensor fusion
Project Metrics
High accuracy sentiment scoring
Multi-approach validation
Technologies Used
BERTHugging FaceNLTKTF-IDFWord2VecSQLite
Project Details
Company
Flunkey
Duration
8 months
Team Size
4 engineers + 2 automotive specialists
Category
NLP