Base Camp
Noida, India
Open to remote & on-site opportunities
Trail Log
Full Stack Developer
Cloud Analogy
Assistant to Lead Developer, delivering full-stack SaaS end-to-end for global product & B2B clients — from PostgreSQL/MongoDB schemas to CI/CD on AWS and Azure — including multi-tenant B2B platforms, real-time systems, and enterprise data pipelines at scale, with a strong AI/LLM focus on production RAG pipelines and MCP tooling.
Key Achievements
- Built and shipped 10 full-stack applications (React/Next.js + Node.js/FastAPI) for 10 client organizations, owning features from database design through production deployment
- Architected multi-tenant, real-time, and enterprise data-sync platforms — including RAG/AI systems — handling large-scale throughput and compliance-grade requirements
- Optimized query and retrieval performance across analytics- and AI-heavy workloads, cutting latency and infrastructure cost by 15% through caching, parallelization, and instance tuning
- Built CI/CD pipelines with review → staging → production promotion, one-click releases, and safe rollback on AWS and Azure
- Mentored 5 junior developers on React/Next.js and code-review practices; served as assistant to the lead across concurrent client engagements
Mountain Range
Each peak represents mastery level — the higher the summit, the deeper the expertise
Trail Map
HOVER OVER WAYPOINTS TO VIEW TRAIL NOTES
Expeditions
AI-Powered Knowledge Assistant — Enterprise RAG Platform
Architect / Lead Developer. End-to-end RAG platform letting organizations chat with their own document corpus via natural language with inline citations.
- ▸Architected an end-to-end RAG pipeline (ingestion → chunking → embedding → indexing → retrieval → answer synthesis) with inline citations back to source records.
- ▸Ran vector search at scale on both Pinecone and Postgres/pgvector, with hybrid semantic + BM25 keyword retrieval for high recall on domain-specific jargon.
- ▸Built a streaming React chat UI with token-by-token rendering; cut median query latency from 8s to under 2s via Redis caching and parallelized retrieval.
- ▸Built a RAGAS evaluation harness measuring faithfulness and answer relevance, so AI features shipped on measured quality rather than guesswork.
Enterprise Data Sync Platform — React Admin Console at Scale
Lead Engineer. Real-time bidirectional sync platform giving a Fortune-500 client a queryable mirror of their CRM data.
- ▸Built a real-time bidirectional sync platform — 10,000+ record syncs/hour across a 40+ object mapping schema with explicit conflict-resolution rules.
- ▸Built a React admin console over high-volume data with server-side pagination, filtering, and virtualized rendering.
- ▸Deployed into network-isolated private spaces with encrypted, audit-logged storage for compliance; ran review → staging → production with one-click releases and safe rollback.
- ▸Integrated 7+ third-party services (monitoring, logging, error tracking, email, caching, scheduled jobs); mentored 3 junior developers.
Field-Service Operations Platform — Dispatch Console + Mobile
Full-Stack Lead. Data-dense operations dashboard and offline-first mobile client for live job tracking and technician dispatch.
- ▸Built a data-dense React operations dashboard for live job tracking, technician assignment, and SLA reporting, with filtering, status overlays, and drill-down detail panels.
- ▸Streamed live position and job-status updates over WebSockets with Redis pub/sub, keeping multiple server instances in sync.
- ▸Shipped an offline-first React Native client with local queueing and conflict-resolved sync for remote, low-signal field sites.
- ▸Tuned instance sizing and concurrency for a 60% infrastructure cost reduction; integrated APM and log aggregation for production observability.
DocuMind — AI Document Intelligence Platform
Personal project. Full-stack platform for structured data extraction from PDF, DOCX, XLSX, and CSV files via LLMs.
- ▸Built a pluggable multi-provider LLM adapter (Groq, OpenAI, Anthropic) users switch per extraction, plus a sandboxed engine for user-defined JS/Python transformations.
- ▸Engineered an async job pipeline (Redis + ARQ) with real-time status over WebSockets.
- ▸Exposed extraction results as MCP tools via FastMCP for direct use in Claude Desktop and Cursor; containerized all services with Docker Compose.
- ▸Implemented an OCR pipeline (PyMuPDF + Tesseract), MinIO S3-compatible storage, JWT auth with refresh tokens, and a team/workspace permissions system.
ChannelHub — Multi-Tenant B2B Reseller & Channel Ops Platform
Multi-tenant REST + Socket.IO platform supporting quotes, opportunities, contracts, sales orders, invoices, and licensing across isolated tenant environments.
- ▸Designed a multi-tenant REST + Socket.IO API across isolated tenant environments.
- ▸Integrated Chargebee billing with HMAC-verified webhook ingestion, DB sync, org-level license reconciliation, and real-time subscription/feature payloads.
- ▸Built async processing on Azure Service Bus with a dedicated queue processor for long-running jobs, relaying live progress over WebSockets.
- ▸Added production observability (Winston logging, Datadog dd-trace) and security patterns (Clerk, JWT, MFA, raw-body webhooks, Zod validation; Swagger docs).
SoundAuth
Passwordless auth platform. Built passkey/WebAuthn authentication with a secure Node.js API and responsive frontend.
- ▸Spearheaded development of a secure passkey-based passwordless authentication system focused on user experience and performance.
- ▸Designed and implemented responsive, modern front-end interfaces while managing scalable API development in Node.js.
Threat Insight
Cybersecurity threat-intelligence platform. Led development of a real-time dashboard UI for high-volume threat visualization and analysis.
- ▸Led design and development of an intuitive, secure UI enabling real-time visualization and analysis of cybersecurity threats.
- ▸Optimized data-rendering pipelines for high-volume event streams, ensuring fast load times and seamless performance.
Trail Badges
Hover or tap badges to flip and reveal details
Trail Companions
“Pardeep is the kind of engineer who makes the whole team better. He owns problems end-to-end — from the first commit to production deployment — and consistently delivers clean, performant code that holds up under real-world load. One of the most reliable devs I've worked with.”
Aniket Singh
Senior Developer @ Cloud Analogy
“What sets Pardeep apart is the rare mix of technical depth and genuine ownership. He can architect a cloud system, optimize a database, and ship a polished UI — all in the same sprint. That level of versatility is exactly what modern teams need.”
Abhishek Kumar
Senior Developer @ Cloud Analogy
“Pardeep doesn't wait to be told what to build — he asks the right questions first. Every feature he touched shipped faster, ran cleaner, and felt more intuitive than scoped. That kind of proactive engineering instinct is incredibly rare and valuable.”
Rutuparna Sakalkale
Product Manager @ Cloud Analogy
“Working alongside Pardeep is a masterclass in full-stack ownership. His code is thorough, his PRs are well-documented, and he never cuts corners on quality. The Multitask Maestro Award he received wasn't a surprise — it was overdue.”
Ankush Saini
Senior Developer @ Cloud Analogy
Signal Fire
Ready to Build Together?
Got a challenging product, an exciting role, or just want to talk tech? I respond fast and I'm always up for a great conversation.