Oct 2023 – Aug 2024 · 10 mos · Bengaluru, Karnataka, India · Hybrid

• Built and optimized APIs using FastAPI and Flask, improving data processing and ensuring reliability.
• Gained expertise in instruments, indicators, issuers, and strategies to support financial analysis.
• Led the implementation of a unified authentication and authorization framework, strengthening security and standardizing access control.
• Refactored the backend to support fund-of-funds reporting, aligning with 12-Factor App principles.
• Enhanced scalability and maintainability for proprietary algorithm-driven portfolio analysis.
• Developed data pipelines with quantitative teams for fund reports aligned to philanthropic preferences.
• Fine-tuned LLMs to automate commentary, reducing manual effort and enhancing insights.
• Worked extensively on real-time financial data infrastructure and reporting for complex investment structures, ensuring accuracy and actionable insights.

Jul 2021 – Sep 2023 · 2 yrs 2 mos · Bangalore Urban, Karnataka, India · Hybrid

• Contributed to the development of multi-threaded, highly available and resilient platforms for ITSM data management using Angular, Python Flask, PostgreSQL, and private cloud (AWS). Developed high-level designs (HLD), low-level designs (LLD), and code, adhering to SOLID principles and 12 factor app standards.
• Led mission-critical Python Scripting and the Go2Cloud initiative for cloud migration to AWS (private cloud), deploying Docker, Kubernetes, Jenkins, Sonar, and other tools.
• Developed a highly modular framework with monitoring using Elastic Search and Kibana, structured logging, alerting, data source connection handling, asynchronous execution, and standardized alerts that accelerates the building of ETL pipelines by 60%.
• Designed and developed 2 crucial data platforms with a Python (Django) backend coupled to PostgreSQL, JWT (User Auth), DRF (REST APIs), and Celery. Led the entire development process, from requirements translation to micro-services-based design, and integrated CI/CD.
• Recipient of the Star of the quarter award ( SOQ ) for rapid, single-handed development of a highly• scalable ( ~10M records/week ) security scan data browsing application besides smoothly regulating core functions.
• Certified 'Data Steward' - Data Citizen programme, conducted by Big Data Academy, SGGSC India and Romania
• Winner of GTS Innovation award 2023 for ideating a phishing evasion solution
• Contributed as an innov-actor to drive and own innovation related efforts across the department
• Actively participated in code reviews, mentored multiple teammates, and made cross-team contributions, including requirement gathering and end-to-end tool development. Contributed to enhancement initiatives to improve ML models’ performance and make ML pipelines more robust and scalable.

Jan 2021 – Jul 2021 · 6 mos

• RCP Crawler:
• Contributed to the development of a distributed cookie management system for a web crawler platform using Python and Java. Developed high-level designs (HLD), low-level designs (LLD), and code for the platform, using Redis, Pub-Sub, Apache Zookeeper, and Apache Kafka.
• Innovated by introducing Playwright instead of Puppeteer (JavaScript crawling) to address dependency management issues. Employed advanced Docker and Kubernetes usage.
• Interacted with stakeholders and took on the responsibility of tweaking and patching the crawler platform involving 5+ micro-services, including orchestration and deduplication modules, to cater to internet scale. Used Java 8, Spring, Kafka, PubSub, Redis, MySQL, and BigTable.
• Revamped the Redis usage, including distributed Bloomfilter (Redis implementation) access and storage, by aiding migration to a managed version.
• Brainstormed a transformed, highly modular architecture for the current platform to address key bottleneck issues, such as IPC (tried gRPC, HTTP implementations), dead letters, and others.

Mar 2020 – Jan 2021 · 10 mos

• RCP Crawler:
• Worked as a key project resource for a highly scalable and resilient platform that crawls over 100+ TB of data per day.
• Developed and implemented robust CI/CD processes for the platform to ensure continuous delivery of features and improvements.
• Possess hands-on experience with various backend development elements such as Concurrent Programming, Lambdas, and Streams, and have used Concurrency concepts such as Phasers, Circuit breakers, Rate limiters, Semaphores, and ThreadPoolExecutor to optimize performance.
• Applied the expertise in performance-oriented technologies like BloomFilters (with), Redis, Kafka, and GCP PubSub to ensure fast and efficient data crawling.
• Developed a structured logging library using Apache Log4J2 as the base, featuring class-level flexibility, JSON logging. Involved artifacting into JFrog for organization-wide use. Used OOP concepts, ThreadLocal-based development, Circular Buffer, Mapped Diagnostic Context (MDC), Factory, and Singleton design patterns to achieve optimal functionality.
• Served as a key resource for the end-to-end development and integration of a captcha solver service (Java + Spring) for the crawler platform using Kafka for asynchronous processing, OpenCV for media preprocessing, and Tesseract (Google’s OS OCR engine) for OCR, achieving an accuracy rate of around 94%. Included HLD and LLD with abstraction and pluggability, without compromising scale.

Jul 2019 – Mar 2020 · 8 mos

• RCP Tools and Services:
• Developed data portability module using Python (Pandas, NLTK) and Django for the annotation platform.
• Developed highly available KPI dashboard APIs with unit testing (Junit) for the annotation platform using Java, SpringBoot, and MySQL.
• Created asset management system for the annotation platform using ExpressJs (NodeJs), RabbitMQ, GCS, and Redis to handle approx. 1TB/day media files.
• Developed a Python script using Beautiful Soup 4 and Puppeteer to crawl data from an E-commerce website using dynamic cookies, resulting in a SPOT award for outstanding contributions across the function.
• Contributed as a core developer of deep learning (CNN + bi LSTM) based image and text classification module with active and continuous learning, in collaboration with Rakuten Institute of Technology. Used MLOps technologies including Tensorflow, Django, and GCP. The technology was submitted for patent registration.

Jan 2019 – Jul 2019 · 6 mos

• RCP Content Ops:
• Developed an attribute extraction platform using Python, Django, Tensorflow, Keras, Fasttext/Word2Vec,
• leveraging NLTK and RegEx. Contributed to end-to-end development, from requirements gathering to deployment.
• Contributed to end-to-end development of a progress tracker tool, utilizing Django, HTML, CSS, JS, and Bootstrap 4. Aimed at aiding daily catalogue operations.
• Developed deep learning models for NER and POS tagging, facilitating attribute extraction.