Responsible for improving Walmart Search by understanding the search queries made by our 300 million+ monthly active customers to search and recommend 1 billion+ products, thereby impacting our half a trillion+ dollar annual revenue

Was responsible for improving Lowe’s search experience by understanding 3 billion+ search queries made by our Lowe’s customers on the digital platforms to personalize/recommend 5 million+ products and impact our ∼100 Billion USD revenue annually. Collaborating across teams to deliver high-performance and low-latency models• Created a Query Classification model for multi-class classification using TensorFlow, PySpark, TensorFlow Serving(TFX) & FastAPI with low latency throughput for search ranking.• Finetuned LLM for Retrieval Augmented Generation (RAG) based on user search query using LangChain with low-latency serving using vLLM and DeepSpeed-MII• Trained a domain-specific BERT to have an accurate & efficient representation of queries and documents related to the home-improvement domain on over 20 million open-source (Semantic Web 2020, Amazon 2018, etc) & in-house data points. The language model outperforms distilled BERT with 20% fewer parameters improving latency.• Conceptualised and built a behavioural tagging model that maps attributes between products and user queries for the top business-driving queries by mining relationships over millions of data points using PySpark.• Expanded search query autocomplete candidate list by 90%+, resulting in more contextual, seasonal and personalised autocomplete suggestions.• Worked on enriching our Elastic Index with user behaviour signals resulting in 16% better recall• Created a 1000 times faster lexical matcher based on SymSpell to compute the importance of each token in a user query for over 3 million + unigrams and bigrams• Created a real-time low latency Spanish Translation model for non-English search queriesSkills: Recommendation Systems, Natural Language Processing, Information retrieval, Deep Learning, System designTechnologies: Tensorflow, PyTorch, Transformers, BERT, Elasticsearch, PySpark, ONNX, FastAPI, GCP, VertexAI

Worked on implementing new features for the JustDial app with 100+ million downloads. Built the very first JD Social (app homepage) content ingestion and recommendation system.• Created a ChatBot with functionalities to serve the user on all 72 verticals (Chatting, Flights, Restaurants, General Search, etc.) that JustDial works on using RASA, SPACY, ChatterBot, Tensorflow, DBPedia• Built a recommendation system for JD Social (JustDial social network) using collaborative filtering and item-based filtering• Created content-based monetizable suggestions. (Eg. suggestion to book tickets to a destination covered in a travel article or buy a product discussed in the article) using BERT (Transfer Learning) with f1 score of 0.88• Summarising news articles using sequence rank to reduce content consumption time.• Researched ways to do Natural Language Processing on Indian LanguagesSkills: Natural Language Processing(NLP), ChatBot, Data Structures, System designTechnologies: Tensorflow, scikit-learn, RASA, BERT, Pandas, NumPy, Flask

Automating processes using Python libraries such as Pandas, OpenPyXl, Keras, etc. and Automation Anywhere

Software development for Power Curve generation using:JSP, Java,Bootstrap and MySQL by performing data analysis on Sensor data