DIV-1 · Divanshu Sharma

Work / Fig. 01 · Founding engineering · Full-stack AI product · 2026

Uniiq — AI Student-Advising Platform

40+ critical vulnerabilities — resolved in inherited legacy code, including prompt-injection and data-leak risks.

Fig. 01

Uniiq — AI Student-Advising Platform

Founding engineering · Full-stack AI product · 2026

Problem

Student onboarding and admissions planning span incomplete profiles, follow-up questions, institution data, and long-running AI calls. The product needed a coherent intake experience without letting partial profiles, transient model failures, or inconsistent admin data leak into downstream workflows.

Approach

Led a technical turnaround across the AWS-deployed React, Express, MongoDB, and Gemini stack: built a three-phase conversational intake with persisted transcripts; added prompt and response sanitization, structured output validation, user-ID rate limiting, transient-503 retries, and recoverable fallbacks; enforced role-aware profile completion; normalized institution and opportunity admin workflows; and code-split 21 routes with explicit Lighthouse budgets.

Results

  • 40+ critical vulnerabilities

    resolved in inherited legacy code, including prompt-injection and data-leak risks

  • 73 → 93 performance score

    LCP cut from 3.3s to 0.7s; INP reduced to 130ms

  • 1,199 lines of tests

    automated coverage added across 12 test modules

  • 21 routes code-split

    PR target: main entry 1.45 MB → 278 KB (81 KB gzipped); not a production measurement

React · TypeScript · Express · MongoDB · Gemini

receipt · figures from Uniiq engineering work summary (2026-07) — private codebase, not publicly verifiable

Uniiq ↗

Release v3.0 · May 2026 — present

Founding Engineer — Uniiq

  • Led the technical turnaround of inherited legacy code, resolving 40+ critical vulnerabilities and adding prompt and response sanitization against injection and data-leak risks.
  • Reworked student onboarding into a three-phase AI conversation with follow-up questions, persisted transcripts, structured validation, rate limits, and recoverable AI-failure handling.
  • Implemented role-aware profile-completion enforcement and international phone capture across protected APIs, sign-up, setup, and profile flows.
  • Expanded institution and opportunity administration with deep-partial updates, nested Mongo normalization, validation, and external-ID uniqueness safeguards; added 1,199 lines of automated tests across 12 modules.
  • Raised the web performance score from 73 to 93, cut LCP from 3.3s to 0.7s, and reduced INP to 130ms after code-splitting 21 routes and optimizing delivery.