“I have worked with Yuvraj first when he interned with me and then joined back next year as full-time employee. In all my interactions I have found him to be very self-motivated and committed towards his work. He always looked for innovative ways of solving tough problems and cared deeply about the quality of his deliverables. I would definitely rate Yuvraj as one of my best hiring decisions. He will be a great asset to any company and I personally look forward to working with him again.”
About
If you're a founder in the Agentic AI or Dev AI space, I offer mentorship and seed…
Experience
Education
Volunteer Experience
-
Volunteer
Pratigya
- 2 years 8 months
Education
Pratigya is a society at Thapar University, Where college students teach under-privileged students after college hours.
I was also honored as the convener of the society in 2013.
Courses
-
Algorithms for Data Science
590D
-
Applied Information Retrieval
590R
-
Computing for Data Analysis (Coursera - Johns Hopkins University)
-
-
Distributed Operating Systems
COMPSCI 677
-
Getting and Cleaning Data (Coursera - Johns Hopkins University)
-
-
Machine Learning (Coursera - Stanford University)
-
-
Neural Networks - Deep Learning
682
-
Probabilistic Graphical Models
COMPSCI 688
-
Social Network Analysis (Coursera - University of Michigan)
-
-
The Data Scientist’s Toolbox (Coursera - Johns Hopkins University)
-
-
Web Intelligence and Big Data (Coursera - IIT Delhi)
-
Projects
-
UMLS entity recognition and linking for biomedical journals
- Present
▪️ Performing entity recognition (identifying diseases and concepts) in biomedical journals and linking (normalizing) the found entities to it's type using Deep Learning and NLP techniques.
▪️ Our proposed Bi-LSTM model for exact entity match in the entity recognition (NER) led to a 3% gain in the F1 score compared to our baseline, TaggerOne model.
▪️ Future work includes building a deep learning model for entity linking and joint optimization of the two tasks. -
Subintent Discovery
-
Trained high quality machine learnt models for mining the hidden actions based on the user query and the entity within. Successfully shipped models for 30+ categories in 2014. Introduced temporal features into the algorithm which improved the overall precision from 91% to 93% with no loss in recall
-
Drug(Medicine) Classifier
-
This project involved training a classifier that would take in a query and classify queries as whether the intent is to find information regarding some drug/medicine or not. I was able to achieve a Precision of 92% and a Recall of 91% on the model that I trained.
-
Data Visualization Layer
-
This project required data (Queries and their related information) to be mined from the Bing repository and use of compelling visualization elements to generate useful insights from the data. This was particularly helpful in analyzing the query patterns, inter-dependence of various classifiers and in identifying important issues like potential false negatives, potential false positives, unidentified entities, ambiguous queries etc.
Honors & Awards
-
Microsoft //Oneweek Hackathon - India Winner
Microsoft
Among the top 5 winning teams of the Microsoft //Oneweek Hackathon Indian leg.
Also received honorable mention in the Microsoft Bing! wide //Oneweek Hackathon (effectively in top 9 projects among 400+ projects). -
ACM ICPC 2011,2012 Regional finalist
-
Qualified for regional finals of ACM ICPC 2012, Kharagpur
Qualified for regional finals of ACM ICPC 2012, Kanpur
Qualified for regional finals of ACM ICPC 2011, Kanpur -
Others
-
- Received Merit Scholarship in the third academic year.
- In Aranya 2011, (technical fest of Thapar University) my team got second place each in Reverse Gear, Crypt Raiders and Virtual Warriors competitions.
- Stood first in a C debugging competition, “C-Combat” and Web development competition organized by Thapar Polytechnic College in 2009.
-
Runners up : Machine Learning Competition
Microsoft STCI India
My team finished second in a Machine Learning competition organised by Microsoft STCI team, India. For the competition, an extracted dataset was provided based on which predictions were to be made. We initially cleaned the data using R, did feature extraction and tried a number of different models using R and Weka Tool. Our final submission used an ensemble classifier which bagged us second position in the competition.
Recommendations received
1 person has recommended Yuvraj
Join now to viewOther similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Yuvraj Singla
-
Yuvraj Singla
-
Yuvraj Singla
Finance Student at the Fordham Gabelli School of Business
-
Yuvraj Singla
Second year Finance and International Business Student at University of Newcastle, Australia
-
Yuvraj Singla
Management Trainee @Kotak Mahindra Bank|| Imt Gzb 25
31 others named Yuvraj Singla are on LinkedIn
See others named Yuvraj Singla