Nov 2024 – Sep 2025 · 10 mos · Bengaluru, Karnataka, India · Hybrid

• Enabled remote upgrades to manage OS-level unattended-upgrades across edge appliances with secure blacklisting of sensitive packages (e.g., NVIDIA, Docker) and improved package consistency across fleet.
• GraphQL APIs, Django models, centralized S3-hosted logging
• Developed asynchronous alerting mechanisms using Opensearch, Celery, and SendGrid to notify admins of appliance health changes, with opt-in user subscription models and user controlled rate limiting.
• Created a security chatbot with fine-tuned GPT-3.5o using openAI assistants API with self built security questionnaire datasets; improved answer accuracy to 95% through dataset cleansing and prompt engineering
• Designed and deployed camera snapshot comparison service using Gemini 2.5 Flash lite through vertexAI APIs to detect visual anomalies (e.g., obstruction, poor lighting) and trigger alerts via Slack.
• Led migration from Zendesk API to Plain.com API for customer support ticketing, including Slack-integrated notifications and automated ticket creation from health events, from appliances and pacs integrations

Dec 2022 – Oct 2024 · 1 yr 10 mos

• Fast Track Promotion in 2 years (minimum requirement is 2.5 years) due to exceptional performance.
• Out of 70,000 employees under our CTO spread in 6 LOBs, only 2500 get to work in the Mission Critical Database Technologies LOB.
• AIM (Automatic In-Memory)
• Oracle caches tables in memory for upto 20X faster OLAP queries. AIM automates this caching end-to-end using some complex caching algorithms.
• Built a simulator using random number generators utilizing various probabalistic distributions to generate workloads which are as close to a real life workload as possible. Implemented LRU, LFU and self developed complex caching algorithms to measure and optimise performance for different workloads.
• ADO (Automatic Data Optimisation)
• Oracle compresses old data to conserve storage. ADO handles this end-to-end. Worked with customers to resolve compression issues, resulting in successful compression and leading to upto 40% storage savings
• Heatmaps
• Oracle has a feature heatmap that tracks accesses to segments and blocks. Optimised the Oracle Block Heatmap API leading to 26X performance improvements for block heat map query times for tables containing millions of rows which are more than 10 GB in size.

Nov 2020 – Dec 2022 · 2 yrs 1 mo

• Customer Flash Cache Size Estimation
• Oracle caches disk blocks in RAM for faster OLTP queries. This cache can be supplemented with a Flash memory cache as well. Estimated the size of flash cache the customer needed to achieve 90% of queries being processed through cache. Analysed the query patterns and accordingly calculated minimum flash cache size requirements.
• Optimised Oracle AIM according to Customer Requirements
• Built an algorithm to optimise OLAP queries by caching tables in RAM, for workloads which have new tables being created after random time intervals.

Aug 2020 – Oct 2020 · 2 mos · Bengaluru, Karnataka, India

• - Got selected in the top LOB of JP Morgan for tech candidates, i.e. GTI(Global Technology Infra)

May 2019 – Jul 2019 · 2 mos · Bengaluru Area, India