Deependra Vishwakarma · Senior software engineer · Mumbai (UTC+5:30)
I build systems that hold.
Senior engineer for systems that have to hold. I design, build and scale web platforms and AI systems end to end — architecture, code, cloud and security — for teams in the US, Europe, the Gulf, Japan and Korea. Based in Mumbai (UTC+5:30), working your hours.
Based in Mumbai (UTC+5:30), working your hours.
Follow one request
- 01
Edge and front end
Next.js, React and Vue front ends served from the edge: fast first paint, accessible, and readable by search engines and AI assistants.
- 02
Gateway
Authentication, authorization, rate limits and input validation: the security that stands in front of everything else.
- 03
Services
Laravel, Node.js and Python services with clear API contracts, tests, and integrations that retry instead of failing silently.
- 04
Cache
Redis and CDN caching, so most requests never reach the database and a traffic spike costs little.
- 05
Queue
Queues and workers for slow work (writes, emails, imports, AI calls), so a surge becomes a longer queue rather than a slower site.
- 06
Database
PostgreSQL and MySQL schemas designed for the next stage of growth, with read paths that scale and backups that have been restored.
- 07
Cloud and delivery
Google Cloud and AWS as code, CI/CD, monitoring and alerts: deployments become routine and incidents become rare.
Round trip complete: the response is back where it started.
A campaign that couldn't go down
I architected and built a real-time voting platform for an influencer awards campaign: a React front end and services on Google Cloud Run behind a Google Cloud load balancer, Supabase for data, and two to three layers of caching in front of it, so the site stayed fast at peak without paying for idle servers.
Hold for 22 seconds, one for each day of the campaign. Space or Enter works too.
Stage 1 of 5
Voting opens. Traffic arrives at the gateway, which checks every request.
1.6M+
users over the campaign
How this is measured
Total users of the voting platform over the campaign, not concurrent sessions. Window: 22-day campaign.
2,000+
concurrent users at peak
How this is measured
Users on the platform at the same moment, at the busiest point of the campaign.
99.99%
availability over the campaign
How this is measured
Availability over the campaign, from the Google Cloud load balancer's health checks: no failed check and no outage was seen during the campaign.
Selected work
- 2024 · Real-time voting platform
InfluencerX
A voting platform that stayed fast for all 22 days of a campaign. - 2024 · CMS migration and modernization
TopGear India CMS
A magazine CMS moved from CodeIgniter on shared hosting to Laravel on Google Cloud: page loads from 1–3 min → 3 s. - 2025 · AI pipeline
Digital Wardrobe
A pipeline that turns a phone photo of clothing into a clean, e-commerce-style product image in 3–8 s. - 2024 · B2B SaaS platform
Exhibit Social
A B2B influencer-marketing platform built from zero: 10,000+ influencers indexed and 30+ brand campaigns run on it. - 2025 · Education platform
Vayudoot Aviations
One platform for an aviation academy: public site, online exams with automatic grading and certificates, and student management. - 2022 · Telemedicine platform
Immunomate
A telemedicine app with real-time video consultations, launched to 500+ registered users.
The AI layer
I add AI to real products: LLM features and retrieval over your documents, agents that act in your tools with permissions and audit logs, and the evaluations and guardrails that keep them honest and within budget. Every engagement starts small, with a one-week scoping sprint on your real data.
AI features in your product
LLM features that work on your real data, measured before they ship.AI agents and MCP servers
Agents that act in your tools with permissions, audit logs and a person in the loop.LLM evals, guardrails and cost control
Know whether your AI feature is getting better or worse, and what it costs.
Your timezone
| Market | City | Overlap, local time | Work week |
|---|---|---|---|
| United States | New York EDT | Mon–Fri 09:00–14:29; Mon–Fri 14:30–18:00 | Mon–Fri |
| United States | Chicago CDT | Mon–Fri 09:00–13:29; Mon–Fri 13:30–18:00 | Mon–Fri |
| United States | San Francisco PDT | Mon–Fri 09:00–11:29; Mon–Fri 11:30–18:00 | Mon–Fri |
| United Kingdom | London BST | Mon–Fri 09:00–18:00 | Mon–Fri |
| European Union | Berlin CEST | Mon–Fri 09:00–18:00 | Mon–Fri |
| United Arab Emirates | Dubai GST | Mon–Fri 09:00–18:00 | Mon–Fri |
| Saudi Arabia | Riyadh GMT+3 | Mon–Thu, Sun 09:00–18:00 | Sun–Thu |
| Japan | Tokyo GMT+9 | Mon–Fri 09:00–18:00 | Mon–Fri |
| South Korea | Seoul GMT+9 | Mon–Fri 09:00–18:00 | Mon–Fri |
Start with a small, fixed-price step.
An audit, an assessment or a scoping sprint, with a written outcome. It's useful on its own, whatever you decide next.