Transforming ideas into robust, scalable solutions through innovative backend development, data engineering, and ML optimization
Passionate about building scalable systems and solving complex problems
Senior Software Engineer with 3+ years of experience at Apple and Intel, specializing in backend development, data engineering, and ML optimization.
I have a proven track record in building scalable distributed systems, processing large-scale data pipelines, and optimizing high-performance computing systems.
My journey includes diverse freelance experiences that built a strong foundation for continuous learning and professional growth.
A timeline of my career progression and achievements
Working on backend development and data engineering for Apple Maps, contributing to data pipelines and authentication systems.
Implemented mutual TLS authentication system for service accounts with client certificate validation during TLS handshake, service identity extraction from certificate extensions, and automatic certificate reloading every 30 minutes without interruption, enabling secure system-to-system authentication for 100+ service interactions
Implemented dual-table Cassandra storage with optimized partitioning strategy for database-level analytics filtering, using atomic batch writes for cross-table consistency, reducing downstream pipeline latency by 40% and eliminating full table scan overhead
Resolved production API rate limiting through pre-caching optimization, reducing backend database calls by 80% and enabling 1000+ concurrent bulk operations without throttling
Engineered end-to-end photo processing pipelines, automating URL discovery, photo download, privacy protection, and quality checks across 50+ vendors, processing 2M+ daily photos and 370K+ features to enhance millions of map locations
Engineered Drive Collection ETL pipeline processing 5TB+ GPS and sensor data daily from global mapping fleet, tracking drive metrics and validating 95%+ data quality with Snowflake integration, and developed global analytics dashboard in Tableau visualizing drive volume, route coverage, and data quality across regions for data-driven mission planning
Led migration of 20+ data pipelines from legacy Hive infrastructure to modern tech stack (Spark/EMR/Snowflake/Iceberg), implementing batch-streaming patterns and reducing data latency by 90%
Worked on optimizing deep learning kernels on Intel Habana Gaudi platforms for AI workloads.
Optimized critical deep learning kernels on Intel Habana Gaudi platforms, achieving 80x speedup for reduction operations, 8x improvement for sparse_segment_sum in object detection models, and 1.5x acceleration for scatter-gather operations that directly enhanced BERT, ResNet, and U-Net performance
Engineered deep learning operators (scatter, gather, maxpool, reduction, softmax) with cross-framework compatibility across PyTorch, TensorFlow, and ONNX, implementing support for multiple precision formats (float32, bfloat16, half, fp8) to enable flexible performance/accuracy tradeoffs on Habana hardware
Designed and implemented the Shared Layer Interface from scratch, significantly reducing model graph compilation time through layer caching, while applying GPU programming with SYCL and OneAPI toolkit for optimal performance across Habana's Gaudi platforms
Building diverse skills through various projects and collaborations
Resolved 50+ student queries across various CS topics with high accuracy.
Taught a client Data Structures and Algorithms in Java and programming fundamentals.
Developed Scala-based Employee Management REST API using Play framework and MySQL.
Developed React.js and Firebase-based admin interface for a farmer-centric application. Enhanced information accessibility on crop diseases, seeds, and fertilizers.
Authored 10+ technical articles on Data Structures and Algorithms in C++.
Led development of a coding interview preparation website using HTML, CSS, and Bootstrap. Created comprehensive learning resources with techniques and code samples.
Designed and optimized 50+ algorithmic problems in C, C++, and Python. Developed solutions for various difficulty levels.
Contributing to the community and developing leadership skills
Community building and technical evangelism for GeeksforGeeks platform. Promoted coding culture and organized technical events.
Student leadership and academic coordination role during first year of B.Tech program. Represented class interests and facilitated communication between students and faculty.
Technologies and tools I work with to build amazing solutions
Contributing to the academic community through research and publications
Implemented geometric algorithm for Moore-Penrose inverse computation of rank-deficient matrices, addressing singularity and non-square dimension limitations of traditional inversion. Utilized range space, null space, and orthogonal projections with LU/QR decomposition to derive novel formula achieving SVD-level accuracy with reduced computational cost.
Comprehensive analysis of COVID-19's economic impact on Indian businesses, focusing on recovery strategies and business continuity planning.
Building a strong foundation in computer science and engineering
Let's connect and discuss opportunities