AI · ML · Backend · Startups

Building AI-powered products that matter.

Hi, I'm Shreyas Tiwary — AI/ML engineer, Co-Founder of Tranquil Labs, patent holder in Federated Learning, and NASSCOM Award Winner based in India.

Passionate founder & engineer

Innovative and research-driven — building AI systems that respect privacy and scale globally.

Recognition

🏆 Awards & Honors

🥇
Best Student Startup in India 2025
NASSCOM & Cisco · April 2025
💰
₹5 Lakh Seed Grant
Tranquil Labs · Hackathon → Commercial
🌐
Global Stage
Startup Mahakumbh · AI Impact Summit · Kerala Global Huddle

I constantly try to improve

My tech stack

Python Flask Docker
PostgreSQL Flutter AWS/Azure

Education

🎓 VIT Vellore

B.Tech Computer Science · 2021–2025

ML · Distributed Systems · Cloud Computing · DSA · DBMS

The Inside Scoop

Currently building Tranquil Labs — AI-powered mental wellness platform.

Do you want to start a project together?

A small selection of recent projects

🧘

Tranquil AI

AI-powered mental wellness platform with CBT therapy, guided meditation, mood journaling & predictive algorithms. 5K+ downloads, 5★ on App Store.

ChatAstro

AI-powered conversational astrology platform delivering personalized astrological insights with prompt-engineered generative AI and precise planetary computations.

🦅

PhoenixHub CRM

Full-stack CRM system with end-to-end lead tracking, automated business workflows, multi-tenant API architecture, and role-based permission tiers.

Next.js SaaS Boilerplate

Production-ready SaaS starter with auth, Stripe payments, dashboard CI/CD — built to eliminate repetitive project setup for every new startup idea.

My work experience

🧘

Tranquil Labs

Current

Co-Founder & CFO

April 2024 – Present · India

Spearheaded technical architecture for an award-winning AI mental wellness platform integrating conversational AI, CBT modules, and predictive mood algorithms. Recognized by NASSCOM & Cisco as India's Best Student Startup 2025.

  • Engineered high-performance backend: Python (Flask), PostgreSQL, Redis, Firebase — containerized via Docker
  • Orchestrated cloud-native SOA across distributed AWS and Azure environments
  • Integrated transformer-based inference pipelines + async Celery/Redis job queues for conversational AI
  • Secured ₹5 Lakh seed funding through financial forecasting models and cloud cost optimization
PythonFlaskPostgreSQLRedisDockerAWSAzureFlutterFirebase
💹

Pravaratan Technologies

Fintech

Software Engineering Intern

Oct 2023 – Dec 2023 · Remote

Contributed to full-stack fintech applications with real-time data pipelines and concurrent transactional systems.

  • Built fintech apps: Flutter frontend + Flask backend with WebSocket-based real-time pipelines
  • Engineered concurrent PostgreSQL transactional systems under high concurrency
FlutterFlaskWebSocketsPostgreSQLFirebase
⚙️

Backstage Army

Backend

Backend Engineering Intern

Sep 2023 – Dec 2023 · India

Managed and optimized backend infrastructure powering the platform's subscription and revenue models.

  • Engineered secure payment tracking workflows for accurate billing and seamless user upgrades
  • Optimized database operations for financial data integrity under concurrent access
BackendPostgreSQLPayment SystemsAPIs
🌐

Tech Head — GDSC

Google Developer Student Clubs · led workshops on cloud, AI & full-stack

🍎

Senior Core — ADG

Apple Developers Group · directed mobile dev events and initiatives

🎓

Senior Core — VITMAS

Organized large-scale academic & technical events at VIT

Research & Intellectual Property

🇮🇳 Indian Patent Application No: 202541122148 · Published 2025

Federated Learning for Privacy-Preserving Mental Health Diagnosis

Proposed and architected a decentralized federated learning framework enabling collaborative AI-driven mental health diagnosis without exposing or centralizing raw patient data. Designed robust secure aggregation mechanisms to mitigate gradient leakage and protect sensitive healthcare datasets from model inversion attacks.

🔒

Privacy by Design

Symmetric encryption on gradient updates aggregated on a central parameter server — raw data never leaves the device

Distributed Training

Local client models compute gradient updates independently, enabling collaborative learning without data centralization

🏥

Regulatory Compliant

Architecture adheres to privacy-by-design principles — viable for real-world medical AI regulatory compliance

📊

Performance Trade-offs

Evaluated convergence speed, network communication overhead, and cryptographic privacy guarantees

Federated LearningPrivacy-Preserving MLSecure AggregationHealthcare AIGradient EncryptionModel Inversion Defense

My approach

+
+
+
+

Architecture First

Before writing a single line of code, I map out system architecture — data flows, API contracts, scalability requirements, and security boundaries to ensure the foundation is solid.

+
+
+
+

Build for Scale

Every system I build is designed to handle orders of magnitude more load — distributed architecture, caching layers, async queues, and cloud-native deployments from day one.

+
+
+
+

Ship & Iterate

I believe in getting things live fast and iterating based on real feedback — from hackathon prototype to NASSCOM award-winning product with thousands of active users.