👨‍💻 Software Engineering
At Blue Altair.
🌏📍India
Connect with me!
[github] [linkedin] [calendar] [biolink]
Check out my blog archive.
I'm a backend engineer with over 4.5 years of experience in full-stack development, specializing in Go and Node.js. My recent work has centered on DevOps/MLOps practices, optimizing deployments and infrastructure. I've contributed to diverse projects in IoT/Building Automation, Digital Media Platforms, and Asset Management, consistently prioritizing clean, maintainable, and scalable architectures that drive real impact.
I believe the greatest bottlenecks in software development are rarely found in the codebase. They exist in the gap between vague business intent and actionable technical requirements.
Having operated as both a co-founder and software engineer, I view my role as a force multiplier. I am a high-agency builder who thrives when given clear, outcomes-based objectives and the autonomy to execute them. My philosophy is rooted in the belief that an engineer’s value is maximized when they are not merely receiving tasks, but understanding the underlying business costs and trade-offs.
In past projects, such as when navigating ambiguous scopes with non-technical stakeholders, I have found that the most effective way to protect engineering velocity is to stop the clock and ask for clarity. Whether I am acting as a translator between commercial stakeholders and engineering teams or building out scalable infrastructure, my priority is the same: document the constraints, define the trade-offs, and ensure the goal is actionable while building minimal prototypes to iteratively receive feedback on.
I bring this same pragmatism to the way I organize work. I advocate for 'pull-based' workflows, where we treat engineers as high-agency adults. By focusing on the importance and impact of a feature (I like to call it the 'Why') I like to foster an environment where engineers choose their work based on the value they can drive, rather than feeling like cogs in a task-pushing machine. This shift in responsibility, from micromanagement to empowerment, produces higher-quality output and significantly reduces the friction that typically stalls high-growth initiatives.
My approach to architecture mirrors this focus on efficiency: I rely on DRY principles and Locality of Behavior to keep systems maintainable, and I prioritize clear documentation and templates to eliminate repetitive overhead. I strive to build systems that are easy to maintain, scale, and understand, so I can keep focusing on the next critical goal.
I'm at my best when I'm solving complex problems in an environment that trades unnecessary process for high autonomy and a relentless focus on outcomes.
At Blue Altair, I developed and deployed MLOps workflows on Databricks (MLFlow), covering feature preparation, model training and production-grade inference. I also designed and implemented cross-system business process automation using platforms such as Workato and n8n.
At Velocilabs, I designed and built the core backend infrastructure for a social media platform, Spyke, using GraphQL, Go, PostgreSQL, and other technologies, which boosted data fetching efficiency by 40%, enabling the frontend team to iterate rapidly and deliver features faster. I also slashed operational costs by 50% through cloud strategy optimization and upgraded deployment strategies for reactive scaling and peak performance.
At Spintly I built APIs for their Building Automation and IoT platform. I refactored core microservices using Saga patterns to improve stability and cut customer complaints by 33%. I also developed and maintained Jenkins CI/CD pipelines to automate the containerization and deployment of microservices. This effort reduced server and infrastructure errors by an impressive 66% over 8 months.
FOCUSA Remind (2025): An open-source Telegram and Slack bot for creating and managing personal reminders, including support for recurring schedules using RRULE. It uses natural language commands (e.g., "Remind me to buy milk at 6pm" or "Remind me to call mom every Monday at 9am") to schedule reminders, which are delivered via the bots. The project is built with Next.js 15, Drizzle ORM for PostgreSQL database management, and Grammy for Telegram bot functionality. It's hosted on Vercel and integrates Vercel's AI SDK for enhanced processing. GitHub
The Green I (2024): A geo-location app that rewards users for supporting local eco-friendly businesses, and also manages rewards through QR codes. Built the backend with Next.js serverless and NeonDB (PostgreSQL). View Project
OGPScraper (2023): A lightweight server written in Go that scrapes OpenGraph metadata (like titles, descriptions, images, and URLs) from websites. It includes caching functionality to handle repeat requests efficiently, reducing redundant scraping. GitHub
DearJarvis (2020): A web application developed for the HackByTheBeach2020 hackathon to help users assess their mental wellbeing through journal analysis. It integrates with note-taking apps like Google Keep to scan and analyze journal entries, providing insights such as mood graphs over time, factors that cause happiness or sadness, desired activities, and tasks to complete. The app emphasizes a seamless writing experience while offering organized insights from past notes. GitHub
My technical understanding covers (non-exhaustive list):
I'm always evolving — currently exploring geo-location application development, and Apigee for API management.
In addition to my professional endeavors, I have also been an independent researcher who has published multiple papers. Collaborating with NIT Goa, we delved into the design, analysis, and applications of ternary logic gates. We also published a paper on High Capacity Conduits (HCCs) and Interface Bottlenecking, both available to read on ResearchGate.
You can connect with me directly on Telegram or LinkedIn.