Dirshak Deepak Patro

I am a final year B.Tech student studying Artificial Intelligence and Data Science, with a strong academic record (9.956/10 CGPA) and an Honours specialization in Data Engineering. I am passionate about systems engineering, backend development, distributed systems, payment infrastructure, and applied AI/ML.

I enjoy building reliable, end-to-end systems and solving engineering problems from first principles. My work includes contributing an empirical-Bayes cold-start routing enhancement to Juspay’s open-source Rust Decision Engine, where I developed a Bayesian scoring approach for sparse payment segments and a reproducible evaluation harness. In synthetic testing, the approach reduced routing regret by 71.7% in sparse segments. I also built DirshakDB, a database storage engine from scratch featuring B+Tree indexing, write-ahead logging, crash recovery, ACID transactions, and an LRU buffer pool.

My other projects include WhatsNextAI, an AI-powered repository intelligence platform using code dependency graphs and retrieval-augmented generation, and BaatSeBharat, a market-regime analysis pipeline combining NLP with financial time-series data.

Through internships at L&T Technology Services and other organizations, I have worked on real-time data pipelines, monitoring dashboards, computer-vision ETL systems, and automation. I have experience with Rust, Python, C, TypeScript, SQL, Redis, ClickHouse, PostgreSQL, Kafka, Apache Spark, FastAPI, and machine-learning frameworks.

I have also been a two-time Smart India Hackathon National Semi-Finalist and co-authored research on deep-learning approaches to ocular disease detection.

I am interested in opportunities in software engineering, backend and infrastructure engineering, distributed systems, data engineering, and AI/ML engineering. I particularly enjoy roles where I can combine strong engineering fundamentals, measurable experimentation, and practical problem-solving to build systems that are reliable, scalable, and useful in the real world.