- TypeScript
- JavaScript (ES6+)
- HTML5
- CSS3
- Python
- Java
Dehradun, India
Somil Sharma
Frontend Engineer
Frontend engineer (React, Next.js, TypeScript) who builds 0→1 products at startup pace. I joined a US startup at ~30 engineers and shipped through its scale-up past 120, building its CRM from the ground up while working daily with US-based product, design, and GTM teams. Before that I built payment flows processing ~65% of revenue at a fintech. AI-native workflow, and I own web performance end to end.
Skills
What I work with
React, Next.js, and TypeScript day to day, plus enough backend and tooling to ship a feature end to end without waiting on anyone.
- React
- Next.js
- Redux
- Responsive design
- Core Web Vitals
- CrUX
- Lighthouse
- SEO
- Node.js
- NestJS
- REST APIs
- Git
- Docker
- AWS
- Mixpanel
- Claude Code
- Jira
- Linear
Experience
Work I have done so far
I started with internships, picked up product ownership early, and ended up building 0→1 products and flows with real business impact.
Aug 2025 - Present
Frontend Engineer at Keychain
- Core engineer on Keychain's CRM, built from scratch: shipped the Accounts and Contacts modules end to end, including list views with search, filter, and sort, detail pages, and activity timelines mapping user relationships against a 100K+ company graph.
- Built inline editing across the CRM's data tables (Accounts, Deals, Connections) with Excel-like cell editing, optimistic updates, rollback on failure, and full error handling.
- Built automated Core Web Vitals monitoring (LCP, INP, CLS) on CrUX data with scheduled Node.js reporting and regression alerting, and used it to catch a production CLS regression.
- Created a high-fidelity, resettable product demo environment for customer demos, with vertical-specific datasets, embedded routing, and feature-flag controls.
- Work AI-first: maintain the team's Claude Code skills and claude.md conventions, and ran AI-driven spec validation across PRD, Gherkin, and Figma docs that surfaced 31 discrepancies before build.
- Frequent first responder on P0 production incidents (broken table scrolling, missing activity timestamps, filter outages), resolving most within 24-48 hours, and ship across the stack in the NestJS backend when features need it.
- Developed the platform's image-cropping flow, now used in ~40% of all image uploads.
- Mentor two interns on SEO initiatives; rated an exceptional performer in the latest review cycle.
May 2024 - Aug 2025
Software Development Engineer 1 - Frontend at smallcase
- Prime contributor to the payments modal on a leading retail-investment platform, powering checkout for ~65% of company revenue by integrating multiple microservices into a seamless, robust payment flow.
- Refactored complex legacy codebases and added automations, improving performance, readability, and developer productivity while reducing technical debt.
- Recognized among the organization's leading performers of the year.
Jan 2024 - Apr 2024
Frontend Engineer Intern at smallcase
- Designed scalable Next.js architecture for the B2B business, and improved page-load times through code splitting, lazy loading, and caching.
Oct 2023 - Jan 2024
Frontend Engineer Intern at Euromonitor International
- Built a ticket-management application in React.js; rated a top performer with a cumulative score of 34/35.
May 2023 - Nov 2023
Founding Engineering Intern at Drone AI
- Developed responsive React.js web applications as one of the first engineers on the team.
Education
Where I studied
A physics background at BITS Pilani, a computer science master's, and certifications picked up along the way.
Master's
Woolf University
- MS, Computer Science
Integrated Bachelor's & Master's
BITS Pilani
- Physics (Theoretical & Computational)
Certifications & Test Scores
Credentials
- Meta Frontend & App Professional Certification (12 months)
- Scaler Frontend Certification - ranked 1st in a batch of 150 students
- TOEFL iBT: 118/120 (2026)
Writing
Publications & Packages
Things I have written and shipped publicly — articles on frontend craft and open-source tooling built from real production problems.
Publications
Preprint · 2026
Accessibility and Semantic Quality Regressions in AI-Assisted React Development: An Empirical Study — Zenodo
Empirical study measuring how AI coding tools (Cursor, Copilot) affect accessibility violations and code complexity in React/TypeScript projects. Uses difference-in-differences with Tobit regression across axe-core audits and AST analysis. Key finding: axe violations show no statistically significant increase attributable to AI assistance; document structure violations are the predominant category.
Packages
npm · Open Source
qualm-a11y
Static AST-level analyzer for LLM-generated React/TypeScript code. Implements the empirical accessibility quality taxonomy from Sharma (2026). Supports SARIF and JSON output for CI/CD pipelines, with a --research-mode flag aligned to the study's violation categories.
1.0.2Version
MITLicense
Node ≥18Engine
qualm-a11yCLI
Build
How I built this site
- Next.js 14
- React 18
- TypeScript
- App Router
- Custom CSS
- Responsive layout
- Semantic HTML
- ESLint
- Static generation
- Production build checks
- Lighthouse audit
What the build looks like
92.2 kBFirst-load JS
6.8 kBHome route JS
85.4 kBShared JS
2.4 kBRoute CSS
4.8 kBResume asset
StaticRendering
Latest Lighthouse results
100Performance
100Accessibility
96Best Practices
100SEO
- First Contentful Paint: 0.8 s
- Largest Contentful Paint: 1.7 s
- Speed Index: 0.8 s
- Total Blocking Time: 0 ms
- Cumulative Layout Shift: 0
Contact
Open to good frontend work.
Based in Dehradun, India, working with US-based teams. Reach out if you are hiring for frontend roles or just want to talk product and engineering.