Jatin Chawla

Hii Visitor 👋
Jatin is currently working for Microsoft's AI Platform and helping Customers with their AI Solutions to accelerate their business transformation.
He aspires to use AI to impact billions of people on the planet.

Work Experience

Microsoft (Bangalore, India) September 2021 - Present

Technical Support Engineer (AI Intelligence Team)
- Solved around 320 customer issues (with 60 super-positive customer verbatims) in helped customers with their AI Solutions. Facilitated customers in shipping their AI Solutions by addressing their maintenance, advisory and technical needs
- Trained my team of 15 engineers on managed online endpoints and collaborated in escalation triages to work on complex live-site customer issues. Triage collaborations helped in speeding up the solution delivery and case closures for 60% of the issues
- Fixed certain parts of the public documentation for the AI platform that helped in lowering the incident counts
- Received Pulse Impact Award, for my efforts to help 20+ customers reduce their code size by almost 40% in terms of volume by helping them levarage functional programming
- Worked with customers to bridge their requirements as feature requests for our AI platform and helped the Product Team with use-case analysis and detailed feature reports
- Apart from my core work, I have hosted and contributed for sessions with Women at Microsoft, Disability Groups and the Aspire (New Hire) Communities. Have been an active evangelist of using AI in technology and a lead-member of the fun club core-committee in my business unit
- Organized the Aspire (New Hires) Stay-Strong Event in India across the 3 sites of Bangalore, Hyderabad and Noida.

IIM Ahmedabad August 2021 - May 2022

ML Research Intern
- Scraped 1 lakh projects from Indiegogo. Post cleaning and tuning the data, finalized 15 important features and trained a semi-supervised Guided-LDA Model that achieved 76% accuracy and almost 80% precision in predicting project-funding success based on the creator’s profile and their language (10 language features) used
- Scrapped 6000 reddits from the science handle to analyze user engagement on those reddits. Performed feature processing to keep 6 independent features (author’s fame, flair, and language features like centered-embeddings, capitalization, using jargons, numeric and special character values) in any post and used Decision Trees to predict engagement with 70% precision

NTU Singapore May 2021 - July 2021

Research Intern
Crafted a hybrid DL model with multiple CNN, LSTM, GRU Cells leveraging attention for classifying medical records into 10 classes. Managed the ML pipeline in a group of 3 by division of tasks. Got an F1-score of 0.75

Tracomo Camera Systems October 2020 - January 2021

Product Development Intern
- Led the development of Social Distancing Software that used a fine-tuned Yolov4-tiny (to accommodate 60 FPS) with a mean squared distance as a threshold implemented (the algorithm gave 95% precision with 3% false positives) with a flask backend and a dashboard for 6 parallel cameras for a chemical industry premise
- Led the development of a Mask Detection algorithm using 4000 images, deployed it with a flask backend and a 6-camera dashboard which was overall 92% accurate in detecting different kinds of masks

Raaga Food Products June 2020 - September 2020

ML Research Intern
- Worked upon Diabetic Retinopathy grading and trained an image model with 10,000 images involving 5 classes
- Used CLAHE followed by median-filtering to preprocess the images that improved the accuracy by 70% to 85%
- Performed ROI cropping to reduce the image size by around 20% that reduced the training time by almost 30% followed by increasing the dataset size to 20,000 images to get a further 4% hike in the accuracy and reached a score of 89%

Projects

Manipulative GAN

- Worked upon a Hybrid Encoder-Decoder GAN architecture to accommodate color manipulation over 30% area in bird images with respect to 5 colors and consisting of 10K images

Wheat-Head Detection

- Fine-tuned a Yolo-V4 with wheat head images from a Kaggle competition and obtained 75% precision in detecting wheat heads with 10% false positive detections initially
- Worked upon ROI for the images that reduced the image size by 15% and doubled the dataset using image augmentation that improved the model performance and took it to 85% and reduced the false positives by 4%

Education

Bachelors in Computer Science and Engineering 2017 - 2021

LJ Institute of Engineering and Technology

- Founder of a Technical club where I led a team of 15 people to handle marketing, design, research, content. The team was responsible to organize events related to programming, placements, personal brand build-up, hackathons. Brought in 2 startups through this initiative to hire 6 interns from the cohort of 400 students

- Courses: Project Management, Algorithms, Data Structures, Computer Architecture, Operating System, Machine Learning, Front End Development, Database Management, Distributed Computing, Advanced Mathematics, Trigonometry and Linear Algebra

Skills

Languages: Python, C++, JavaScript, MySQL

Tech Stack: Machine Learning, HTML, CSS, NodeJS, Flask, Azure, NLP

Tools: JIRA, FIGMA, VS-Code, Tableau, PowerBI

Others: Product Management, Public Speaking

Contact

If you want to connect with me for any projects, collabs, or just to say a 'Hi' please feel free to mail me at chawlaj00@gmail.com

LinkedIn

GitHub