- Validated REST API responses for enterprise web modules, catching response-schema mismatches and edge-case failures during QA.
- Fixed frontend and backend defects in JavaScript and SQL — resolving UI rendering issues and incorrect query outputs across features.
- Documented 20+ defects across QA cycles, tracking resolution and coordinating fixes with developers using Git and Agile sprints.
OPEN TO SWE & AI/ML ROLES · 2027
Building intelligent systems across AI, software & the web.
Tiya Agarwal — CS undergraduate at VIT Vellore.
I build everything from federated learning research and AI-powered platforms to production-style full-stack applications, RAG systems, AIOps dashboards, and embedded TinyML devices.
02
Software Engineering Internships
12
Projects Built
200+
DSA Problems Solved
04+
Hackathons & Challenges
01 — Experience
Two internships spent finding what breaks.
Both roles put me deep in QA and defect resolution — which taught me to think about software the way testers do: what fails, what's missed, what silently returns the wrong answer.
- Wrote SQL queries to validate migrated business records, resolving data inconsistencies across inventory and accounting modules in TallyPrime.
- Tested ERP inventory and accounting workflows end-to-end, logging defects and verifying data integrity post-fix.
- Engineered Excel-based reporting templates to track migration progress and flag discrepancy patterns during ERP go-live.
02 — Featured work
Twelve projects across research, systems & the web.
From a federated-learning framework for pediatric oncology to an AIOps platform and an on-device fitness wearable. Click any project to open the full case study.
03 — Security
Security work: find it, prove it, fix it.
Beyond building systems, I test how they break. My VAPT work is hands-on and evidence-driven — real exploitation, CVSS scoring, and verified remediation, not just naming OWASP categories.
Black-Box VAPT Engagement
NoteVault
A self-contained penetration test: a deliberately vulnerable Node/Express app, a real black-box assessment against it, a professional report with proof-of-concept exploits and CVSS 3.1 scoring, and a remediated build with automated retest evidence proving each fix.
CRIT · 9.1SQL InjectionUNION dump of credential store
HIGH · 8.7Stored XSScookie theft via note comments
CRIT · 8.1IDORcross-user note access by ID
CRIT · 9.1Broken Authforged JWT via hardcoded secret
HIGH · 7.5Misconfigurationunauth debug endpoint leaks secret
04 — Technical ecosystem
The stack I build with.
Hover a domain to light up the tools I reach for. Everything here is drawn from shipped work.
05 — About
I like understanding how systems work — then building something from it.
I enjoy working across different layers of technology. Some projects start as research questions around federated learning and machine learning; others become full-stack products, RAG systems, AIOps platforms, or embedded devices.
What connects them is the same instinct: I like taking apart how something works and then building something tangible from that understanding. That runs from a class-imbalance-aware aggregation rule for distributed model training, to a <50ms inference loop on a microcontroller, to the unglamorous QA work that keeps a product honest.
Right now my focus sits at the intersection of AI/ML, backend systems, full-stack development, generative AI, and distributed systems — and I'm looking for roles where I can keep learning and building at that intersection.
06 — Certifications
Credentials
- AWS Certified Cloud PractitionerAmazon Web Services · CLF-C02View ↗2026
- Machine LearningSmartBridge × Google for Developers · CC-ML-2025-13799View ↗2025
- Cyber Security — Industrial TrainingEdu-versity (MSME) · EDU10815View ↗2023
- Pygame Workshop — ISTEgraVITas'23, VIT VelloreView ↗2023
- Journey to Cloud: Envisioning Your SolutionIBM SkillsBuild2025
- Getting Started with Artificial IntelligenceIBM SkillsBuild2025
07 — Beyond the code
Achievements & leadership
200+
DSA problems solved
Across Easy, Medium & Hard on LeetCode.
4+
Hackathons & challenges
Rapid-prototyping ML and web solutions end-to-end under time pressure.
50+
IEEE Computer Society · Core Member
Led AI/ML research discussions and peer-learning sessions.
PR
Public Relations Head · VIT Dance Team
Led event outreach, branding & publicity campaigns.
08 — Contact
Let's build something interesting.
I'm always interested in hard engineering problems, ambitious products, and opportunities to learn and build. The fastest way to reach me is email.