Sakil Ansari

AI-POWERED MARKET INTELLIGENCE AND ADVANCED Text based FORECASTING

Developing an AI/ML based digital solution to automate the forecasting tools/technique.The digital product/proposed solution is able to make supply and demand predictions for PE(Polyethylene) in China with accuracy close to what manual efforts result and able to spot demand anomalies of different types and break them down by segments, and into specific value drivers causing them.

NAOMI-Virtual Assistant

Naomi is a virtual assistant for order management in booking order.Developed Table Extraction System from PO documents.The virtual Assistant saves the time for booking order by OP agent.Used FasterRCNN for table line-item detection, CRAFT for text-extraction.

Capacity Planning

Manual process for order Forecasting was automated with ML/DL technique.The system is helping to manage the resources for a paticular quarter and helps to generate the revenue by providing the number of resources which are required for the business for a certain quarter.

Carrier Online Products Sentimental Analysis

An AI based automated digital solution which extracts the products reviews from various data sources. ML/DL algorithms are developed to find sentiment scores and also topic modeling system is developed . These help the business to enchance the user experience of the products based on the users reviews.

Publications

• Experimental Analysis of the Changes in Speech while Normal Speaking, Walking, Running, and Eating
• Coin Recognition System using Artificial Neural Network on Static Image Dataset
• Facial expression recognition
• Implementation of Magnitude and Phase Spectrum Compensation to Achieve an Enhanced Speech
• Speaker Diarization of Broadcast News audios

Publications

• Automatic Cloth Pattern and Color Recognition for Visually Impaired People Using SVM Algorithm
• TWITTER SENTIMENT ANALYSIS USING MACHINE LEARNING
• Pattern Recognition Techniques:A Review
• Evaluation and Comparative Study of Edge Detection Techniques
• IMAGE QUALITY ASSESSMENT FOR FAKE BIOMETRIC DETECTION: APPLICATION TO FINGERPRINT AND FACE RECOGNITION