Jitender Kumar Email & Phone Number

Utilizing NLP and deep learning techniques in talent acquisition, including named entity recognition(NER),part‐of‐speech(POS)tagging ,and sentiment analysis, to analyze resumes and job descriptions effectively.Contributed to the development and application of AI models, including large language models(LLMs),for tasks such as resume parsing and candidate recommendations. Kept abreast of advancements like retrieval‐augmented generation (RAG) architectures to enhance recruitment workflows.Developed matching algorithm capable of ranking candidates with 80% accuracy within 20 seconds per candidate, significantly reducing recruiters’ time in sorting resumes by 50%.Lead team to create NLP pipelines by fine‐tuning transformer models and constructing question‐answering systems to better align applicant experience with available positions.

Pune, Maharashtra, India

Built model to discover intents,entities and conversation paths to identify and map automation opportunities for bot development/training from the audio transcriptions.Created custom chatbot models for Jaguar(https://www.jaquar.com/en/), Utkarsh(https://utkarsh.com), ViduraX(http://www.vidurax.com) having nlp understanding using TensorFlow.Lead a project in which we built a recommendation system for a dating application.We built persona of each of the profiles by asking them about their interests and asking some AI generated questions.We are also trying to make this as a SaaS service so it can be easily integrated with any other applications which needs persona of peoples(Finding a fitness instructor).Collaborated with WowLabz to create GRA(Global Resource Allocation)for ABInBev.Algorithm allocates the resource using the historic data on different levels and taking sub‐packages defined by user as consideration.application.We are trying to make this as a SaaS service so it can be easily integrated with any other dating applications.

Bengaluru, Karnataka, India

Worked on time series forecasting under guidence of Sudhakar Suga and Ranganath Prabhu.Predicting the trend and able to predict the problem that will occur in component so that precautions can be taken prior to that.We have tried LSTM and HMM but don't have enough data for LSTM and the fluctuation is not read by HMM properly.So used SVR, it gives better results than others but is not good on a long span of time.

Bengaluru, Karnataka, India

Worked as a data scientist to build models for the ZScore website under the guidance of Deleep Murali and Ayush Banka.We have completed following projects :-1. Generic Semantic detection :- Correctly detecting the semantic type of data columns.It is also crucial for data science tasks such as automated data cleaning, schema matching, and data discovery.2. Automatic Data-correction:- A statistical inference model to impute and enrich data having missing values. It leverages available quality rules, value correlations, reference data, and multiple other signals to build a probabilistic model that accurately captures the data generation process.3. Domain Semantic detection :- We took a heuristic approach and build a algorithm of our own.We divided the task in 2 different models.One which will detect the semantic labels out of table and after having the labels we predict the domain by analyzing combination of labels.4. Entities matching from text