Vinay Katiyar

AI Researcher

Gurugram, Haryana, India12 yrs 10 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

12+ years of experience in Data Science
Expertise in NLP and Generative AI
Strong focus on deploying ML models at scale

Stackforce AI infers this person is a Data Science expert specializing in NLP and AI solutions for SaaS applications.

Contact

vinay.has.katiyar@gmail.com LinkedIn

Skills

Core Skills

Natural Language Processing (nlp)Machine LearningDeep LearningStatistical ModelingData Visualization

Other Skills

Anomaly DetectionCNNCollaborative FilteringData MiningData ModelingDecision TreesDjangoEntity ExtractionImage ProcessingK-Means ClusteringKerasLSTMMatlabMultivariate AnalysisNER

About

• 12+ years of experience in Data Science, specializing in NLP, Generative AI, Agentic AI, and ML Engineering with LLMs (GPT, BERT, LLaMA, etc.). • Strong focus on deploying ML/AI models at scale using cloud-native tools (Azure, GCP), Docker, Kubernetes, CI/CD. • Skilled in building intelligent agents, vector search, and ML solutions aligned with user engagement and personalization

Experience

12 yrs 10 mos

Total Experience

2 yrs 6 mos

Average Tenure

4 yrs 9 mos

Current Experience

Jio platforms limited (jpl)

Lead Data Scientist

Sep 2021 – Present · 4 yrs 9 mos · Gurugram, Haryana, India

Resume Parser: Build a tool to parse resume and extract all the information to be filled in candidate’s profile page. This system is based on multiple classifier and entity extraction algorithm. We mainly used RCNN, XGboost, and other rule based classifier to build this tool.
Patent has been submitted on resume parser
Skill Extraction from Resume and JD:
Build a system to identify Tech and Non-Tech skill from text present in resume and jd. This algorithm uses context awareness approach to classify skills. We used mainly CNN model to build this system.
Job Recommendation System:
Build a system which finds best matching candidate for the job. This system process candidate profile including resume and job description. In this tool, we develop an algorithm which uses the extracted entities and create a matching score. System uses various algorithm like Siamese Network similarity, NER, CNN based skill extraction module and finally ensemble based approach to generate final matching score.
Intent based cross-recommendation in News Notification System:
In this tool, we are extracting intent and entities appear on news and social media and used these information to build news-recommendation system. This system is currently under developing phase.
Course Recommendation:
This tool is built to recommend courses to the user. We process the watching history of user, his peers and his career aspirations to recommend various technical and management courses. We Used pyspark to process the data and collaborative filtering based algorithm for model training.

Natural Language Processing (NLP)Machine LearningEntity ExtractionCNNXGBoostSiamese Network+1

Mobileum

Sr Data Scientist

May 2019 – Sep 2021 · 2 yrs 4 mos · Gurgaon, India

Lead analytics and technical innovation through data driven model development using traditional Machine Learning methods
Develop algorithms and build data models for use in various telecom related problems like IoT detection, Audio classification, anomaly detection, and travel prediction
Build ML solution pipelines for feature development and training/retraining for product usability and enhancement.
Leverage innovative solutions to develop customized model output visualization
Develop expertise in Telecom, Bigdata, and other supporting technology relevant to company
Work in cross functional team to deliver innovative data driven solution to different clients across globe
Implemented deep learning based models like RNN/LSTM to model time series data in various telecom related problem
Participate in mentoring junior data scientist and knowledge sharing sessions across verticals

Machine LearningDeep LearningRNNLSTMData ModelingAnomaly Detection

Oneassist consumer solutions

Sr. Data Scientist

Oct 2017 – Apr 2019 · 1 yr 6 mos · Gurgaon, India

Developed KPI based analytical and statistical model to capture the Fraudulent Behaviour of customers based on their historical data
Create CNN based image classifier to identify faulty mobile phone images submitted during policy activation. It helps to optimize labour cost and human error
Build and deployed an Decision Tree based Turn Around Time(TAT) prediction model using Django python API. Capture approximately 95% variance on various steps involved during claim execution
Experience in all phases of analytic process including data collection, preparation, modelling, evaluation, and deployment

Statistical ModelingImage ProcessingDjangoDecision Trees

Exl analytics

Data Scientist

Sep 2016 – Oct 2017 · 1 yr 1 mo · Gurgaon, India

Build segmentation model for Auto Insurance client to find out high risk zone in US. Model is based on K-Means clustering algorithm.
Develop an OCR tool based on Tesseract to extract relevant information like Name, Place, Mobile No., address etc from scanned pdfs and images.
Build and deploy Visualization tool based on SAS and Tableau to understand cost breakdown for insurance client
Communicate with client and develop various report using Pivot-table and different charts like Bar plot, Pie plot, Histograms etc

K-Means ClusteringOCRSASTableauData Visualization

Tata consultancy services

Data Scientist

Jul 2013 – Sep 2016 · 3 yrs 2 mos · Pune

Developed a NLP based tool to extract relevant information from research paper. It is based topic modelling and Document classification. Model clusters different domain related papers and on top of it tree based classifier is developed that provided relevant portion of text
Developed and attrition model for BPO segment to understand and analyse root cause for high attrition rate. Data used is mainly related to employment history, appraisal/appreciation achieved and educational background.
Build an activity mining tool to provide the most productive time of employee. Tool mainly used employees activity data they perform on the office machine which involves text and images. We used various statistical models to develop activity mining tool, it mines the data and predict probable future activity.
Participate in authoring various research papers based on data mining and Agent based modelling.

Natural Language Processing (NLP)Data MiningStatistical Modeling