Software Engineer → Team Lead
Apr 2025 — Mar 2026
Built the backend for an AI-powered edtech platform helping students revise smarter through adaptive learning, RAG-based study assistance, and intelligent quiz generation. Promoted from Software Engineer to Team Lead within the first month.
Led backend architecture as Team Lead: broke features into FRDs approved by founders, assigned and reviewed subtasks via PR review, and introduced a documentation-first workflow that cut average feature delivery time by 50%.
Designed a hierarchical RAG pipeline (parent/child chunking across Chroma and DynamoDB) with exact page-level source citations, wrapped in a 3-layer guardrail system (input, context, output) to catch and prevent hallucinated responses.
Built an adaptive learning engine that scores per-topic student mastery using a forgetting-curve-weighted model, dynamically adjusting explanation difficulty and generating personalized study roadmaps.
Built a hybrid rule-based + LLM question extraction pipeline that parsed past papers into structured, rubric-tagged questions (MCQ, long-form, fill-in-the-blank), powering weakness-weighted mock test generation.
Built the backend for an AI-powered edtech platform helping students revise smarter through adaptive learning, RAG-based study assistance, and intelligent quiz generation. Promoted from Software Engineer to Team Lead within the first month.
Led backend architecture as Team Lead: broke features into FRDs approved by founders, assigned and reviewed subtasks via PR review, and introduced a documentation-first workflow that cut average feature delivery time by 50%.
Designed a hierarchical RAG pipeline (parent/child chunking across Chroma and DynamoDB) with exact page-level source citations, wrapped in a 3-layer guardrail system (input, context, output) to catch and prevent hallucinated responses.
Built an adaptive learning engine that scores per-topic student mastery using a forgetting-curve-weighted model, dynamically adjusting explanation difficulty and generating personalized study roadmaps.
Built a hybrid rule-based + LLM question extraction pipeline that parsed past papers into structured, rubric-tagged questions (MCQ, long-form, fill-in-the-blank), powering weakness-weighted mock test generation.
Software Developer
Oct 2025 — Dec 2025
Solo designed, built, and deployed the backend for a real-time cricket scoring platform supporting tournament management, live scoring, and player statistics across complex match and playoff formats.
Designed and normalized a 31-table PostgreSQL schema with 50+ composite indexes; added denormalized stat tables to cut read load and resolved N+1 queries, reducing API response times from 500ms to 200ms.
Built a real-time scoring system using an event-sourced ball-by-ball log, allowing full match state to be reconstructed from the log after a server restart.
Built contextual RBAC (roles scoped per team, tournament, or match) with JWT authentication, a custom email + OTP flow, and Google/Facebook/Apple OAuth.
Implemented dual-tier rate limiting (global + per-endpoint) to prevent abuse and control OTP delivery costs.
Solo designed, built, and deployed the backend for a real-time cricket scoring platform supporting tournament management, live scoring, and player statistics across complex match and playoff formats.
Designed and normalized a 31-table PostgreSQL schema with 50+ composite indexes; added denormalized stat tables to cut read load and resolved N+1 queries, reducing API response times from 500ms to 200ms.
Built a real-time scoring system using an event-sourced ball-by-ball log, allowing full match state to be reconstructed from the log after a server restart.
Built contextual RBAC (roles scoped per team, tournament, or match) with JWT authentication, a custom email + OTP flow, and Google/Facebook/Apple OAuth.
Implemented dual-tier rate limiting (global + per-endpoint) to prevent abuse and control OTP delivery costs.
Project Trainee
Jan 2025 — Mar 2025
Completed the backend fundamentals portion of a structured engineering training program, working through tasks modeled on real prior production issues within an R&D team.
Worked through tasks modeled on prior production issues, covering schema normalization, indexing, and async task offloading.
Practiced professional code review workflows and REST API design discussion within an R&D engineering team.
Built hands-on experience with structured backend fundamentals, including DTO/Pydantic schema conventions and database design.
Completed the backend fundamentals portion of a structured engineering training program, working through tasks modeled on real prior production issues within an R&D team.
Worked through tasks modeled on prior production issues, covering schema normalization, indexing, and async task offloading.
Practiced professional code review workflows and REST API design discussion within an R&D engineering team.
Built hands-on experience with structured backend fundamentals, including DTO/Pydantic schema conventions and database design.
Web Development Intern
Jul 2024 — Dec 2024
Designed and built a production full-stack eCommerce platform for a custom gifting business (personalized print-on-demand products and time-guaranteed cake delivery), serving live customers. Sole backend engineer after an initial two-month ramp-up.
Built JWT + OAuth (Google/Facebook via NextAuth) authentication with a 3-tier RBAC system (customer, store, admin), including dedicated admin and store control panels.
Designed a dynamic checkout flow handling standard, customizable (image-upload-dependent pricing), and delivery-slot-constrained product types, integrated with Razorpay.
Owned deployment on AWS EC2 with Docker and S3-backed image storage.
Diagnosed and fixed a production page-load performance issue by converting images to WebP, identified using Google PageSpeed.
Designed and built a production full-stack eCommerce platform for a custom gifting business (personalized print-on-demand products and time-guaranteed cake delivery), serving live customers. Sole backend engineer after an initial two-month ramp-up.
Built JWT + OAuth (Google/Facebook via NextAuth) authentication with a 3-tier RBAC system (customer, store, admin), including dedicated admin and store control panels.
Designed a dynamic checkout flow handling standard, customizable (image-upload-dependent pricing), and delivery-slot-constrained product types, integrated with Razorpay.
Owned deployment on AWS EC2 with Docker and S3-backed image storage.
Diagnosed and fixed a production page-load performance issue by converting images to WebP, identified using Google PageSpeed.