Gunjan H. Profile

UDFORSK

Data Mining

Data Processing

Machine Learning (ML)

Natural Language

Statistics

0,0

(0 anmeldelser)

Du følger nu

Fejl under følg af bruger.

Denne bruger har ikke tilladt andre bruger at følge dem.

Du følger allerede denne bruger.

Din medlemsskabsplan tillader kun at følge 0. Opgrader her.

Du følger nu ikke længere

Kunne ikke stoppe med at følge bruger.

Du har nu anbefalet

Kunne ikke anbefale bruger.

Noget gik galt. Forny venligst siden og prøv igen.

Email verificeret.

$5 USD / time

kolkata,

india

$5 USD / time

Det er i øjeblikket 2:20 PM her

Tilmeldt maj 24, 2023

0 Anbefalinger

Gunjan H.

@gunjanh99

0,0

(0 anmeldelser)

0,0

100%

$5 USD / time

kolkata,

india

$5 USD / time

N/A

Jobs færdiggjort

N/A

Indenfor budgettet

N/A

Til tiden

N/A

Genansættelsesrate

Data Scientist | IIT Dhanbad Graduate

I am a professional working in an MNC as an AI Practitioner. Having topped in diverse hackathons ranging from college to international level and a handful of professional internships and fellowships, I aim to apply the knowledge gathered from these experiences in your projects and help develop some state-of-the-art solutions. I have three years of industry experience and have published multiple papers in top AI, ML, and NLP journals. I am proficient in oratory and management skills, which may further aid my work. I can work diligently and with perfection to contribute significantly and live up to your expectations if given a chance. I would be glad to send you any other information and welcome any suggestions on your part. I eagerly look forward to your response. Neughleana#6050 Portfolio: [login to view URL]

Se alle Data Processing Executives i India

Kontakt Gunjan H. omkring dit job

Log ind for at diskutere detaljerne via chat.

Portefølje

This paper presents a novel approach for predicting the path of active particles using time series analysis, LSTM networks, and Transformers. The proposed method outperforms other approaches' accuracy and robustness, especially in complex scenarios with particle interactions and noise.

Brownian Motion of A-Particles Using Pre-LN Transformers

Engineered a library that can automatically detect driving patterns while driving and set the battery parameters to the required states accordingly; to save the battery for longer life using IIoT, Neural Nets, and Statistics. Re-Trained TV4DX-LSTM, based on real-time data from SoC and PI, by 6.4%

PyBMS:BMS using Sync-LSTM with Parallel Rectified Filtration

Multi-Purpose Rank Estimating Recommendation Engine Using Blend of RankNet and LambdaMart with Normalized Discounted Cumulative Gain (NDCG) The pipeline developed a machine learning algorithm to rank the top 93 companies based on customer base, outreach, and social media presence. RankNet and LambdaMart were used for ranking, and NDCG/MAP was used to assess model performance. K-Nearest Neighbors was used for feature engineering.

Rank Estimating Recommendation with RankNet & LambdaMart

SaaS companies are customer-driven and are heavily dependent on their customer base. There are a set of metrics that can showcase the health of the SaaS companies and their aspects. These metrics and numbers are not readily available on publicly reported SEC Filings and need to be chalked out.

SEC Filling Analyzer for SaaS Companies using EDGAR Data

Causality detection draws plenty of attention in the field of Natural Language Processing and linguistics research. It has essential applications in information retrieval, event prediction, question answering, financial analysis, and market research. In this study, we explore several methods to identify and extract cause-effect pairs in financial documents using transformers. For this purpose, we propose an approach that combines POS tagging with the BIO scheme, which can be integrated with modern transformer models address this the challenge of identifying causality in a given text.

Combining POS-tag with Context Embed for Causality Detection

An efficient and robust communication channel is a critical factor in underground coal mines to ensure the safety of the working environment. Applying predictive machine learning of convergence, load at withdrawal and duration of yielding in hrs will automate the analysis of the after-effects of depillaring operations. Trained GLEs and GBEs ensembled using XGBoost and Random Forests Regressor on CTGANs sampled data to predict timed Convergence/Withdrawal Load on Rise/Dip side and achieved R2 score 0.91, 0.89 respectively. Developed an advanced technique to extract the ’yielding time’.

AI-based Predictive Analysis of (SAGES) for Mines Safety

Anmeldelser

Ændringer gemt

Ingen anbefalinger at se her!

Erfaring

Nielsen

jul. 2022 - Nuværende

• Scrapped product review data up to 94% using BeautifulSoup, OAuth2, and REST API to make quality datasets for modelling • Sorted text data based on relevancy using Sorensen-Dice Coeff, Twersky Index with Multihead-Siamese Nets, and SMART-Roberta-Large and applied Text Clustering using G-BAT with 89% efficiency. Performed analytics on Lexical Richness

IISc Bangalore

okt. 2021 - jul. 2022 (9 måneder, 1 dag)

• Engineered a library for Battery Management Systems (BMS), enhancing functionality using IIoT, Neural Nets, and Statistics • Discovered a novel multi-channel driving pattern LSTM network, TV4DX-LSTM, based on real-time data from SoC and Performance Index (PI), improving the SOTA MAPE score by 6.4% upon using the Kalman Filter

Larsen & Toubro

jan. 2021 - sep. 2021 (8 måneder, 1 dag)

• Built a semi-automatic annotation tool for labelling point cloud image data on a large scale relying on multi- sensor approaches • Boosted with advanced pixel gradient algorithms to provide bounding box annotations on probable objects, which include a pointwise strategy to guide the annotator and help manage the dataset with 85% less human intervention

Uddannelse

Bachelor of Technology (IIT Dhanbad)

Indian School of Mines,

India

2019 - 2023

(4 år)

Kvalifikationer

IBM

2021

The IBM Machine Learning Professional Certificate is an online training program by IBM that equips learners with skills in machine learning. It covers Python programming, data science, supervised and unsupervised learning, deep learning, reinforcement learning, and time series analysis. Learners gain hands-on experience with popular libraries and work on a capstone project. The certificate signifies proficiency in machine learning techniques.

TensorFlow - Google Brain

2021

"DeepLearning.AI TensorFlow" is an online course that offers a practical introduction to deep learning using TensorFlow. It covers fundamental concepts, such as neural networks, CNNs, RNNs, and GANs. Learners gain hands-on experience building, training, and evaluating models using TensorFlow. The course emphasizes real-world applications, transfer learning, text data processing, and model deployment in practical scenarios.

University of Michigan

2020

The University of Michigan's course "Applied Data Science with Python" is a comprehensive introduction to data science using Python. It covers topics such as data manipulation, visualization, machine learning, text mining, social network analysis, and includes a capstone project. The course equips learners with the skills needed to analyze, visualize, and model data using Python.

Publikationer

POS tagging with Attention-based Contextual Representations for Identifying Causality in Documents

ACL, FinCausal 2021

Causality detection is important in NLP and linguistics research. It finds applications in information retrieval, event prediction, question answering, financial analysis, and market research. We explore methods for extracting cause-effect pairs in financial documents using transformers. Our approach combines POS tagging with the BIO scheme, integrated with modern transformers. Our best methodology achieves an F1-Score of 0.9551 and Exact Match Score of 0.8777 in the FinCausal-2021 Shared Task.

Artificial Intelligence-Assisted Efficient Analysis of SAGES for Improving Underground Mines Safety

ResearchSquare

Efficient analysis of SAGES in underground coal mines using historical data. Predictive machine learning automates after-effect analysis, aiding maintenance and safety. Linear estimators, clustering methods detect SAGES output. Fine-tuning approach for algorithms compared. Ensemble models used for target variables.

Forecasting Brownian Dynamics of Active Particles Using Pre-LN Transformers with L-LSTM Embeddings

NeurIPS (upcoming)

This paper introduces a novel method for predicting active particle paths using time series analysis, LSTM networks, and Transformers. It outperforms other approaches in accuracy and robustness, especially in complex scenarios. The research provides insights into particle dynamics and contributes to more accurate predictions with broad scientific and engineering implications.

Battery Management System using Driving Pattern Synchronised LSTM with Parallel Rectified Filtration

IEEE

Digital transformation drives innovations in transportation with ICT, AI, and IIoT. Our research focuses on developing a digital twin of BMS for lead acid batteries. PyBMS is proposed as an enhanced BMS digital twin using LSTM network and parallel rectified filtration for real-time SoC and Performance Index estimation. It dynamically identifies driving pattern and battery anomalies, integrating V-EKF algorithm as an enhancement to EKF.

Evaluation and Comparison of Mechanical Subsystems of Mobile Roof Support Systems with SAGES

ResearchSquare

This study objectifies the role of mobile roof support (MRS) systems in the area of underground mining operations to improve stability and worker safety while presenting a comprehensive comparison and evaluation of various mechanical subsystems and operations of mobile roof supports used worldwide with Self Advancing Goad Edge Supports (SAGES) taking in minor consideration of integration of AI and IoT-based sensors to develop appropriate and efficient load bearing architecture.