How I Solve Problems
Real examples of how I identify root causes, redesign systems, and deliver reliable production outcomes
About Myself
I'm a DevOps and backend engineer with about 3 years of hands-on experience across the full delivery lifecycle — from writing backend systems and fixing performance bottlenecks, to building CI/CD pipelines and making on-call survivable. I started as a frontend developer, moved into backend and SRE work, and I'm now full-time DevOps at InfoZ IT Services where I own deployment pipelines, cloud cost optimization, and infrastructure reliability. What I enjoy most is finding the root cause of things — whether it's a broken pipeline, a slow query, or a noisy alert system — and fixing it properly, not just patching it. I tend to own things end-to-end, and I don't consider a problem solved until it stops coming back.
Career Milestones
Rebuilding Broken Deployment Pipelines
Inherited pipelines that regularly broke on releases. Rebuilt them with staged rollouts and automated rollbacks. Deploy failures dropped to near-zero in the first month.
InfoZ IT ServicesCloud Cost Optimization — 40% Waste Found
Audited infrastructure and found 40% of instances were oversized or idle. Right-sized compute and introduced scheduled scaling — cut costs without impacting availability.
InfoZ IT ServicesMaking On-Call Survivable
Replaced vague alerting with structured, threshold-tuned monitoring. MTTD went from user-reported to under 5 minutes.
LDM CollegeBuilding a College ERP from Scratch
Architected a zero-trust, edge-secured Academic ERP (library, attendance, finance) serving 500+ daily users, eliminating dual-fee data anomalies and NoSQL injection vectors.
LDM CollegeExam Submission: 8 Seconds → Under 1 Second
Identified synchronous DB writes as the bottleneck under peak load. Moved to batched async inserts.
LDM CollegeFull Table Scans → Instant Page Loads
Rewrote heavy queries doing full table scans. Added composite indexes and materialized views.
HCL TechnologiesEliminating Recurring SEV-2 Incidents
Inherited a system with 3+ SEV-2 incidents per week. Focused on root causes instead of faster restarts.
HCL Technologies200+ Alerts/Day → Under 20 Actionable
Overhauled alerting rules that had become pure noise. Restored trust in the monitoring system.
Personal ProjectAI CommodityChain — Real-Time Market Intelligence
Built a commodity intelligence platform — live data feeds, local 70B LLM, context-aware analysis, Pearson correlation matrix across global energy and metal markets.
Technical Deep Dives
AI CommodityChain — Real-Time Market Intelligence
Built a commodity intelligence platform combining live price feeds, locally-hosted 70B LLM, context-aware analysis, and a dynamic correlation engine — all with 60-second data latency.
Cloud Infrastructure Automation Platform
IaC solution using Terraform and AWS CDK for provisioning secure, highly available multi-account cloud environments with drift detection and automated rollback.
Enterprise Web Scraper — 10,000+ Products Tracked
Replaced manual, inconsistent competitor data collection with a Scrapy + Selenium pipeline tracking 10,000+ products continuously — structured for real analysis, not raw dumps.
AI Trend Analyzer Pipeline
Automated data pipeline that scrapes tech sources, processes content via custom LLMs to extract emerging trends, and surfaces insights on a real-time dashboard.
Sales Reporting — From Manual Excel to Automated Pipeline
Replaced manual Excel-dependent reporting with a Python pipeline (Pandas + OpenPyXL) delivering consistent, decision-ready reports automatically.
Serverless Microservices Architecture
Scalable backend leveraging AWS Lambda, API Gateway, and DynamoDB with asynchronous event-driven queues and zero-maintenance operations.
Email Marketing — From Batch Blasts to Behavior-Triggered
Replaced fixed-schedule campaigns with a behavior-triggered system using SendGrid API — making campaigns adaptive without increasing manual effort.
Kubernetes CI/CD Deployment Engine
Full CI/CD pipeline with GitHub Actions and ArgoCD. Zero-downtime Blue/Green deployments with automated health checks and rollback triggers.
API Integration Framework — Write Error Handling Once
Built a reusable Flask-based integration layer that standardizes external API connections — handling retries, JSON/XML parsing, and error handling once so every integration inherits reliability by default.
Document Management — Rule-Based Auto-Organization
Built a rule-based file categorization system using regex — replacing unstructured file storage with automatic organization that handles edge cases like conflicts and duplicates explicitly.
Social Media Suite — Centralized, Reliable, Analytical
Replaced fragmented manual social posting with a centralized system using Selenium and platform APIs — with scheduling, cross-platform reliability, and analytics feedback loop.
Experience Quick Reference
| Situation | Angle | Impact |
|---|---|---|
| Broken pipelines at InfoZ | Rebuilt with staged rollouts + auto-rollback | Near-zero deploy failures |
| Cloud overspend | Audited + right-sized + scheduled scaling | 40% instances were waste |
| Noisy monitoring at InfoZ | Threshold-tuned, silenced noise | MTTD: user-reported → 5 min |
| ERP from scratch at LDM | Architected for bursty load | 500+ DAU, still running |
| Exam submissions slow | Async batched DB writes | 8s → under 1s |
| Slow DB queries | Composite indexes + materialized views | Page loads instant |
| SEV-2s at HCL | Root cause analysis, not just restarts | Recurring incidents eliminated |
| Alert noise at HCL | Audit + retune + runbooks | 200+ alerts → under 20 |
| AI CommodityChain | Context injection + local LLM | Grounded, real-time analysis |
Project Quick Reference
| Project | Problem | Decision | Result |
|---|---|---|---|
| AI CommodityChain | Hallucinating AI financial tools | Context injection from live data | Grounded analysis, 60s latency |
| Enterprise Web Scraper | Manual, stale competitor data | Scrapy + Selenium dual-engine | 10,000+ products tracked reliably |
| Sales Dashboard | Manual inconsistent Excel reports | Repeatability-first pipeline | Consistent, decision-ready reports |
| Email Automation | Generic batch campaigns | Event-based triggers + SendGrid | Targeted, adaptable campaigns |
| API Framework | Rebuilding integrations from scratch | Reusable adapter + failure-first | Hours not days per integration |
| Document Management | Unstructured file storage | Rule-based + edge case handling | Consistent, searchable storage |
| Social Media Suite | Fragmented manual posting | API + Selenium + analytics loop | Reliable cross-platform scheduling |