Who I am
Akshay Kaushik
AI Full Stack Engineer · Django · Next.js · 3+ years production
I'm a full stack engineer with a production-first mindset, working at the intersection of Django backends, Next.js frontends, and AI integrations. I've been building with AI tools since GitHub Copilot's early access — not as a trend, but because it genuinely makes you faster and forces you to think about what code is actually for.
My path hasn't been linear. I spent time as an independent engineer building and growing a content platform to 4M+ annual visits — that taught me more about performance, SEO, and product thinking than any formal role could. A Google algorithm update ended it overnight. I kept building.
Right now I'm deepening my AI engineering work — building real retrieval systems, multi-agent pipelines, and AI-native products. Not wrappers. Systems that handle real load and real failure modes.
Work history
Experience
Software Engineer
Apr 2024 – PresentMukul Consultants India Pvt. Ltd. · MrSolvo.com · Noida
- —Built a two-stage AI product search pipeline — Django backend, Google Embedding API for vector similarity, Gemini LLM as a second-pass contextual reranker. Connected to a Next.js frontend.
- —Designed zero-gap invoicing covering 4 transaction types: platform fees, provider payouts, user refunds, and multi-scenario edge cases. Fully audit-ready ledger.
- —Implemented JWT edge authentication on CDN — role-based route protection and post-login redirects for high-traffic routes.
- —Built internal ops console (React Router) covering KYC, payments, refunds, ledger, and bookings with IP-allowlist access.
- —Led technical interviews and onboarded new engineers through their first delivery milestones.
Founder / Independent Engineer
Aug 2021 – Jan 2024Self-directed · Remote
- —Bootstrapped and grew a content platform to 4M+ annual visits from zero — full stack, no team.
- —Handled everything: Core Web Vitals optimization, CDN configuration, technical SEO, PageSpeed tuning, and ad/affiliate monetization.
- —Platform was hit by Google's Helpful Content Update — killed overnight traffic that took years to build. I learned more about algorithmic risk and dependency than any job could have taught.
- —Used GitHub Copilot from its early access days. This was my entry point into AI-assisted development — I've been building with AI tools continuously since.
- —Took freelance web development work throughout this period to stay in production code.
Backend Developer
Jun 2019 – Apr 2020Bdsharma Consulting Pvt. Ltd.
- —Built a document digitization ETL pipeline processing thousands of offline bank inspection reports (PDF and Word) from multiple empanelled banks.
- —Used regex-based extraction to parse report IDs, dates, locations, and bank names across format variants.
- —Loaded structured records into Django backend; async batch jobs via Celery + Apache Spark for volume. Deployed on DigitalOcean.
Context
On the Independent Years
Between August 2021 and January 2024, I wasn't employed at a company. I was building one — or rather, building for myself. I grew a content platform from zero to 4M+ annual visits using modern web fundamentals: Core Web Vitals, CDN architecture, technical SEO, PageSpeed, and ad/affiliate monetization. Then Google's Helpful Content Update hit and eliminated the traffic I'd spent years building.
I don't present this as a gap. I present it as one of the most formative periods of my career. I learned what it means to be accountable to your own systems — no team to catch failures, no manager to escalate to. Just you, the code, and the consequences.
Throughout this period I was also using AI tools actively — GitHub Copilot from its earliest access, then progressively more capable tools like Codex, Claude Code, Cursor as they arrived. That early, hands-on relationship with AI tools is part of why I can work with them effectively today: I didn't adopt them as a trend, I grew with them.
Technical domains
Technology
AI & LLMs
Backend
Frontend
Infrastructure
Right now
What I'm Building
CareerLens — An AI-native career intelligence platform that uses LangGraph multi-agent architectures, GPT-4o, OpenSearch, and reflection loops to help people understand their career trajectory. Production-grade. Real reasoning. Not a chatbot wrapper.
Get in touch
Connect
Open to senior engineering roles, AI consulting, and interesting technical conversations. If you're building something hard, I want to hear about it.