Coact.ai
Flask + Azure OpenAI coaching platform. A FAISS RAG engine adapts to GROW and STAR frameworks; sentiment analytics give real-time feedback across 5+ communication metrics.
- Flask
- Azure OpenAI
- FAISS
- RAG
AI engineer / software developer / backend builder
I'm Suyash Balasubramaniam, a final-year B.Tech AI & Data Science student in Chennai, graduating in 2027 — and a Red Hat Certified System Administrator who turns ambiguous ideas into dependable products, practical AI and data-driven systems.

Selected projects / 2024–2026
Systems built end to end — from database schema to a deployed, evaluated model.
Flask + Azure OpenAI coaching platform. A FAISS RAG engine adapts to GROW and STAR frameworks; sentiment analytics give real-time feedback across 5+ communication metrics.
Thermoelectric system converting transformer waste heat into usable DC electricity. A real-time CMS streams sensor data; predictive maintenance and a ROS2 digital twin are in development.
Django + PostgreSQL lifecycle platform with CRUD, enterprise IT compliance, a complex relational schema and RBAC for 1,000+ business-critical asset records.
NestJS APIs and MySQL service-ticket operations. Firebase Cloud Messaging delivers status updates to Android and cross-platform field technicians.
Software engineering · AI · backend
Five roles across data, applied AI, backend engineering and public-private digital transformation — each one shipping real software.
Part of the national Dx-EDGE initiative (CII · NITI Aayog · AICTE) — trained on the Digital Maturity Diagnostic Framework to drive frontier technology adoption across India's MSME sector.
Built NestJS service-ticket lifecycle modules, MySQL schemas for 1,000+ records, and FCM technician alerts.
Built a Flask + Azure OpenAI coaching platform; shipped a FAISS coaching engine with 95% framework adherence and cut reported inference latency by 25%.
Delivered Django + PostgreSQL asset management with RBAC and secure relational data modelling.
Automated ingestion with Python and UiPath, turning unstructured files into JSON for AI training workflows.
Illustrative system architecture
A view of the shape behind a production-ready AI application: requests are routed, context is retrieved, output is evaluated and every useful event becomes a signal.
The intelligence plane. Retrieval, grounding and evaluation compose context into a useful, verifiable answer.
Illustrative figures — simulated traffic, latency and uptime that show the shape of a production AI application, not live telemetry.
01 Designed for clear boundaries
02 AI is a service, not a black box
03 Observability is part of the product
What I actually build
Applied AI applications, backend services, production engineering and intelligent products — categorized so recruiters can skim fast.
Academic + verified credentials
Engineering foundations, a Red Hat certification, and leadership that ships outside the classroom too.
Sri Sairam Engineering College, Chennai · 2023–2027 · CGPA 7.5
Sri Chaitanya Techno School, Pallavaram, Chennai
Vision International School of Excellence, Ankleshwar, Gujarat
Red Hat Certified System Administrator — practical Linux administration, service configuration and observability.
On campus and off
Technical leadership, event coordination and national-level sport.
Interactive shell / this portfolio
This site ships with a working bash-style terminal. Type help to explore my projects, experience and links the way a systems person would.
Open to Software Engineering · AI Engineering · Backend Engineering