B.Tech in AI & Data Science
2023–2027
AI & DATA SCIENCE ENGINEER • CHENNAI, INDIA
VIGNESH BALAMURUGAN
BUILDING SYSTEMS THAT
REASON, VERIFY & SHIP.
I turn applied AI research into dependable products—from multi-agent workflows and RAG systems to computer vision and production backends.
- 03
- AI internships
- 8.82
- Current CGPA
- 2×
- GATE streams
What I bring
Ideas are cheap.
Evidence is the work.
I am a B.Tech AI & Data Science student at IIIT Sri City who enjoys the difficult middle between a promising model and a system people can actually trust.
My work spans agent orchestration, retrieval and factual verification, vision research, backend architecture, evaluations, and deployment. I care about the unglamorous parts too: validation, failure states, audit trails, reproducible tests, and clear product decisions.
Discuss a role or collaboration ↗From evaluation fixtures and test suites to cloud deployment.
Computer vision and intelligent transportation manuscripts under review.
Qualified in both Computer Science and Data Science & AI.
Work history
Building with
real constraints.
Ownership across AI architecture, backend reliability, and applied research—not isolated notebook demos.
GenAI Engineering Intern / AI Lead
ResonateAI
Leading a seven-member AI roadmap for an autonomous, multi-framework app-generation platform. Built the LangGraph Planner–Coder–Reviewer flow, structured validation, bounded repair, follow-up regeneration, and production queue integration.
- LangGraph
- Django
- Celery
- Redis
- Groq
Backend Lead / AI Full Stack
The AI Signal
Designed Atlas Intelligence's modular backend and reliability layer, including source claims, confidence-based approval, conflict detection, manual review, audit logs, and deployed ingestion flows.
- TypeScript
- Hono
- Prisma
- Neon
- Cloudflare
AI Research Intern
NIT Tiruchirappalli
Engineered a ViT–LSTM human-activity-recognition pipeline for video streams and an end-to-end video-captioning system. Co-authored research currently under review.
- PyTorch
- ViT
- LSTM
- OpenCV
- Video AI
Data Analytics Virtual Internship / Job Simulation
Deloitte via Forage
Completed a practical analytics simulation covering data wrangling, KPI dashboard design, and insight synthesis on unstructured client datasets modeled on Deloitte consulting work.
- Data Analytics
- Data Wrangling
- KPI Dashboards
- Insight Synthesis
Featured builds
Systems with
receipts.
Six projects selected for depth, reliability, and range. Every metric is scoped to its committed evaluation—not inflated into a universal claim.
AgentMesh
A multi-tenant platform for trustworthy customer-support AI, with hybrid retrieval, claim-level verification, safety vetoes, escalation, immutable knowledge releases, and auditable decisions.
AppSpec Compiler
Compiles a natural-language product prompt into validated AppIntent, DataSchema, and AppSpec contracts—with targeted repair and deterministic fallback.
Sentinel PII Redactor
An offline, DOCX-aware pseudonymization tool that replaces sensitive entities consistently across paragraphs, tables, headers, footers, notes, and comments.
OPD Claim Engine
A rule-first, nine-stage insurance adjudication pipeline combining deterministic eligibility checks with medical, fraud, and explainability layers.
VEXHALIA
A multimodal travel-intelligence agent that routes between vector retrieval and live search, then runs weather and image collection concurrently before synthesis.
REVARX
An abandoned-cart recovery system with automated follow-up, tone A/B tests, reply-intent classification, journey timelines, and campaign analytics.
More completed work
Honors research
STRIDE.
Training-free lane discovery & travel-direction estimation for fixed traffic cameras.
STRIDE turns repeated vehicle motion into an interpretable map of road structure. It combines tracking, density and motion-coherence maps, circular statistics, and rule-based region extraction—without training a lane-segmentation model.
The current revision evaluates the method across 47 UA-DETRAC scenes. Public wording here stays deliberately conservative while review is ongoing.
Explore the research repository ↗International Journal of Intelligent Transportation Systems Research
Journal manuscript IJIT-D-26-00620 remains under peer review. Revision work is in progress; this is not an acceptance claim.
JOURNAL TRACK / IJITSRIDICON Conference Submission
The conference version is currently under review. Acceptance and publication details will be added only after an official decision.
CONFERENCE TRACK / IDICONCapability map
Technical skills
built for delivery.
Intelligence
LangGraph · LangChain · MCP · RAG · FAISS · ChromaDB · Hugging Face · Multi-agent systems
Models
PyTorch · TensorFlow · scikit-learn · OpenCV · Vision Transformers · RL · NLI
Product
FastAPI · Flask · Django · Node · React · REST · TypeScript · PostgreSQL · Redis
Operations
Docker · Kubernetes · GitHub Actions · MLflow · DVC · Azure · AWS · Prometheus · Grafana
Languages
Python · Java · C · SQL · JavaScript
GenAI Tooling
LangChain · LangGraph · MCP · AI Agents · Hugging Face Transformers · RAG · FAISS
AI & ML
PyTorch · TensorFlow · scikit-learn · OpenCV · Reinforcement Learning · Vision Transformers
Cloud & DevOps
AWS · Docker · Kubernetes · GitHub Actions · CI/CD · Git · GitHub · Linux · MLflow · DVC · Prometheus · Grafana
Databases
MySQL · MongoDB · PostgreSQL · Redis · ChromaDB
Backend
FastAPI · Flask · Node.js · REST APIs · SQLite
Core CS
DSA · OOP · Operating Systems · Computer Networks · System Design · API Design · DBMS · Distributed Systems
Machine Learning specialization
Advanced Deep Learning specialization
NPTEL Programming in Java
GATE 2026 qualified
Proof of learning
Credentials
you can inspect.
Seven completed learning, simulation, and competition records, each linked directly to the original credential.