Sagar Sharma

I am a Full Stack Developer with a deep ambition to achieve mastery, constantly seeking to refine my skills and deliver exceptional results. I thrive on autonomy and independence, taking full accountability for crafting logical, high-quality solutions that bridge innovative ideals with pragmatic execution.

Experience

U

Full Stack Developer

Uniprep · Jun 2024 – Present

Not yet confirmed

• Engineered the end-to-end architecture of Uniprep's CUET prep platform using React/Next.js, React Native (Expo Router), and Node.js/PostgreSQL backend services, scaling from MVP to 50K+ registered users and 5K+ daily active users.
• Architected a high-concurrency computer-based test (CBT) engine replicating the official NTA user interface with section timers, auto-save, and resume-on-disconnect, sustaining 2K+ concurrent test-takers with sub-200ms write latency.
• Designed and implemented an adaptive practice engine using an Item Response Theory (IRT) difficulty model and student-topic mastery matrix, driving personalized question selection and lifting average mock exam scores by 23% across 10K active learners.
• Built a multi-tenant content management system managing a 250K+ item question bank with granular RBAC, bulk CSV/LaTeX math ingestion, and a 4-stage review workflow, cutting content publishing turnaround by 60%.
• Implemented secure user authentication and payment workflows, including JWT/OTP auth, active device-session management, and Razorpay recurring subscription plans featuring coupon logic, automated webhook reconciliation, and idempotent transaction handling.
• Engineered a real-time analytics engine processing 12M+ attempt events to compute NTA-style score normalization, percentile ranks, accuracy heatmaps, and time-per-question metrics, cutting report latency from 8s to under 400ms via Redis.
• Modeled the core PostgreSQL schema (60+ tables) and optimized performance by implementing composite indexing, connection pooling, and partitioning of the attempts table by exam cycle, keeping query latencies under 200ms.
• Integrated Redis-backed background job queues (BullMQ) to handle heavy asynchronous operations, including scorecard PDF generation, automated email performance reports, study material PYQ packaging, and cohort push notifications via Expo.
• Developed a Delhi University (DU), JNU, and BHU college & course predictor mapping CUET scores to cutoff trends for top colleges like SRCC, Hindu, and Miranda House, integrated with a course preference-filler dashboard.
• Built internal operations consoles for content management, student support tracking, and cohort segmentation, enabling a 15-member academic and support team to manage mock operations independently.
• Configured automated GitHub Actions CI/CD deployment pipelines and optimized the React Native mobile app bundle size via code splitting, lazy loading, and asset caching, reducing initial app load time by 40% on lower-end devices.

Founding Full Stack Engineer

UNIPREP Learning · Jun 2024 – Present

Not yet confirmed

• Working as a Full Stack MERN Developer, building and scaling a CUET preparation platform used by students across India.
• Developed and maintained web applications using React.js and Next.js, mobile applications using React Native, and backend services with Node.js, Express.js, Supabase, and PostgreSQL.
• Built core product features including online mock tests, adaptive practice modules, analytics dashboards, content management systems, authentication and subscription payments, AI-powered college prediction, and internal admin tools.
• Focused on designing scalable backend services, optimizing database performance, improving API response times, and delivering a reliable experience for both learners and internal teams while collaborating closely across the product lifecycle.