Data Scientist - Noida/Bangalore, India - PhysicsWallah

Description

Working Experience : 3 to 7 Years.

Startup experience preferred, Edtech work experience bonus.

Location : Noida & Bangalore (WFO).

Roles & Responsibilities :

- Help teams understand what Data Science can do for them and set the right expectations.

Use deep learning, machine learning and analytical methods to create scalable solutions for business problems.

Create innovative solutions and applications utilising advanced NLP algorithms/architectures including (but not limited to ) LLMs for tasks such as text generation, summarization, translation, entity extraction and concept recognition, clustering, and more.

Contribute to the execution of our vision for NLP-based technology solutions using various NLP toolkits like Huggingface, Spacy, CoreNLP, OpenNLP, etc.

Perform relevant data analysis and benchmark the NLP solutions to improve our offerings.

Be able to clearly communicate results and recommendations to various stakeholders.

Evaluate the effectiveness of the solutions and improve upon them in a continuous manner.

- We expect one to have a mix of a strong technical background, the ability to understand the business implications of their work, and the ability to empathise with our users and work towards helping PhysicsWallah give them the best experience.

Help and mentor junior members to become better data scientists.

Skill Sets :

- 3+ years relevant work experience in building NLP(ML/DL) models.

Strong foundational knowledge in transformers ( BERT, GPTs, T5s etc.), Embeddings.

Expertise in SQL, Python(Must).

Hands-on experience with latest GenAI modelling (GPTs, Mistral, Falcon, LLama) and approaches (RAG, Langchain ) etc.

Experience using machine learning (structured, text, audio/video data) libraries (preferably in Python), deep learning (Tensorflow, PyTorch).

Good to have :

- Foundational knowledge of any cloud (AWS/Azure/GCP).

- Expertise in querying relational, non-relational, graph databases.

Experience in big data technologies (Spark, MapReduce, Pig, Hive).