Headline

Software Engineer | AI/ML Engineer @ Fox Corporation | Backend Systems • Distributed Data Pipelines • LLMs • Search & Recommendation • Databricks | Top 2% CRED CodeChef Hiring Challenge

About

Software Engineer & AI/ML Engineer passionate about building production-grade backend systems, scalable AI applications, and intelligent data platforms. During my internship at Fox Corporation, I worked on production Databricks/PySpark pipelines powering semantic search, recommendation systems, LLM-driven metadata enrichment, embeddings, and large-scale content understanding across a 700K+ media catalog. My interests span Machine Learning, Generative AI, Retrieval-Augmented Generation (RAG), LLMs, Recommendation Systems, Distributed Data Processing, Backend Engineering, and Applied AI. I enjoy designing reliable, scalable systems that bridge software engineering with modern AI. I'm actively seeking Software Engineer, Machine Learning Engineer, Applied AI Engineer, GenAI Engineer, and Applied Scientist opportunities across India and global teams. Core Technologies: Python • Java • SQL • Databricks • PySpark • FastAPI • Apache Spark • Docker • GitHub Actions • PostgreSQL • AWS • GCP • LLMs • RAG • Vector Search • Embeddings

Location

Gurdaspur

AI stack

langchain

langgraph

openai

huggingface

LlamaIndex

PydanticAI

Work experience

AI/ML Engineer Intern - Personalization & Recommendation Team | Databricks Platform at Fox Corporation, Jan 2026 - Jul 2026 · 6 mo, Bengaluru, Karnataka, India, • Built and deployed end-to-end AI/ML pipelines powering semantic search, content discovery, and recommendation across a 700K+ media catalog (shorts, clips, listings, series, movies, and specials) using Databricks, PySpark, LLMs, and vector embeddings. • Engineered LLM-driven metadata enrichment workflows using GPT-5/Gemini to transform ASR transcripts and catalog text into 40+ structured metadata attributes, improving downstream search relevance, ranking, and personalization. • Designed a hash-based Change Data Capture (CDC) framework that processed only modified content, significantly reducing LLM inference costs and eliminating redundant embedding generation. • Built scalable embedding pipelines and optimized distributed processing using Delta MERGE, Spark optimizations, caching, adaptive query execution, and a 300-thread concurrent processing framework, enabling reliable high-throughput production workflows. • Developed and monitored production Databricks pipelines supporting recommendation systems, semantic retrieval, and content-understanding use cases, leveraging telemetry, logging, and batch-level metrics for operational reliability. • Collaborated closely with software engineers, product managers, and ML teams in an Agile environment to debug production issues, review code, and deliver scalable backend data platforms.

Education

Bachelor of Engineering - BE, Computer Science Engineering, Thapar Institute of Engineering & Technology

Public links

leetcode: https://leetcode.com/bhumika_mahajan01

github: https://github.com/BhumikaMahajan87

bhumikamahajan87.github.io: https://bhumikamahajan87.github.io/bhumika-portfolio/portfolio/