Manjunath Karanth

Background

About Manjunath Karanth

Seasoned Python Developer with over 6+ years of experience. Specialist in ETL processes, optimization, and developing automated scripts, backends, and APIs. Focused on NLP, LLM, and Generative AI, having trained and deployed models for various ML tasks. Experienced in containerizing applications using Docker, deploying on GCP, and processing big data with EMR clusters on AWS. Proven ability to identify automation opportunities and develop effective solutions, combining technical skills with creative thinking. Elevated from Senior Creative Associate to Technical Lead, Assistant Director, and now Associate Manager, my journey exemplifies innovation and technical skill across varied domains. Highlights include: Machine Learning, LLM, Gen AI, NLP (Python): •Deployed Paddle OCR to detect text in a large video database and used fine-tuned Llama 2 to identify offensive content. •Trained YOLOv8 models to find inaccuracies in Indian political maps. •Fine-tuned and deployed LLaMA2 for fact-checking and offensive content analysis. •Fine-tuned Hugging Face RoBERTa for sentiment analysis on customer reviews. •Audited sales calls using OpenAI's Whisper for transcription, Docker for scaling, MongoDB for the database, and ChatGPT to detect mis-selling. Data Processing (Python): •Performed numerous ETL processes using Selenium, lxml, pandas, and PySpark on AWS. •Developed custom functions for processing requests sent from mobile devices via Telegram, using a Telegram bot, Flask, Selenium, and MongoDB. Software & Web Development (Python, Flask, HTML): •Created a Python script to automate batch editing and stitching of videos with transitions. •Developed a GUI tool using PyQT for seamless PowerShell script execution across domain PCs. •Built a RAG system for backend and a basic frontend for an internal website for efficient video database search and playback. Automation (JavaScript, Python): •Developed over 25 AFX Tools and 10 Premiere Pro functions for editing efficiency. •Standardized After Effects templates and automated compositing with the Auto Comper tool. •Enhanced Blender workflows and optimized Thinkbox Deadline for parallel computing. Computer Graphics (Adobe): •Standardized templates with expressions and utilized Element 3D and Trapcode for compelling videos. •Excelled in chroma keying and compositing; streamlined processes with custom scripts to reduce production times. •Focused on quality improvement by automating repetitive tasks.