jitendrasinghiitg / divolte-kafka-druid-superset

A proof of concept for Click Stream Collection Process using Divolte, Kafka, Druid and Superset

Jitendra Singh

LinkedIn | GitHub | [email protected] | [phone]

---

Summary

Accomplished Machine Learning Engineer with 7+ years of experience in developing and deploying advanced ML models and algorithms. Expertise in Python, TensorFlow, PyTorch, scikit-learn, and deep learning techniques. Skilled in data analysis, feature engineering, model training, and optimization.

Passionate about leveraging cutting-edge technologies to solve complex problems and drive innovation.

Education

B.Tech in Electronics and Communication Engineering

Indian Institute of Technology, Guwahati | IN
Graduated: May 2017

• Relevant Coursework: Linear Algebra, Probability and Random Processes, Machine Learning, Deep Learning, Digital Signal Processing, Speech Processing, Image Processing

---

Skills

• Programming Languages: Python
• Frameworks: Flask, Fastapi
• Databases: MySQL, PostgreSQL, MongoDB
• Tools: Git, Docker
• Libraries: TensorFlow, PyTorch, Keras, Scikit-Learn, Pandas, NumPy, Open CV
• Cloud Platforms: Google Cloud Platform (GCP), Microsoft Azure, AWS(EC2, S3, Sagemaker)

---

Experience

Senior Machine Learning Engineer

TrueFan | Gurugram, IN
May 2022 - Present

• Enhanced GAN-based lip-generation model, using wave2vec2 features, raising accuracy from 53% to 70%, enabling lip-sync videos with Indian celebrities.
• Quantized GAN model and achieved 2x inference speed and cutting costs to half.
• Implemented a text-to-speech model for voice cloning using Microsoft SpeechT5, reducing dependency for vendor to only 20%.
• Trained Hindi MFA model for better alignments for indian names, outperforming English Models. Thus signifi- cantly boosted Lip-Sync accuracy from 50% to an 75%.

Senior Data Scientist

Decimal Technologies | Gurugram, IN
May 2019 - Jan 2022

• Created OCR for diverse POI and POA documents, Streamlined workflows by reducing manual data entry time by 80%.
• Streamlined Image Analysis by doing Blur Detection, using Laplacian and Text Orientation for Documents in Image, using hough transform, Implementing data validation rules to reject images that do not meet certain criteria at source by 90%
• Developed a PDF parser for bank statements, ITR, and GST certificates, automating data extraction for loan applications. Significantly reduced manual data entry by 90%, streamlining the process.
• Cibil ,Bank Statement, Epf data analysis for helping Lenders to better verify a persons actual income and expen- diture in 90% of the cases.

Junior Data Scientist

Tatras Data | Delhi, IN
Jun 2015 - Apr 2017

• Engineered and deployed a News recommendation system, incorporating advanced data science techniques. Achieved a remarkable 30% increase in user engagement and satisfaction, solidifying the platform’s value proposition and competitive edge.
• Utilized Random Forest, and Gradient Boosting to predict critical health condition for rheumatoid arthritis. Improved junior doctors’ decisions, reducing hospitalization rates by 20% and enhancing overall patient care.
• Designed an online apparel retailer’s recommendation system, applying advanced Machine Learning to develop a hybrid model. elevating customer engagement by 20%, boosting sales, and increasing conversion rates, driving substantial business growth.

---

Certifications

• Earned Generative AI with Large Language Models from coursera
• Earned Deep Learning Specialization from coursera

---

Languages

• English (Conversational)
• Hindi (Native)