Top
Based in Dublin, IE
Local time ··:··
Status Available

Shipping AI Systems That Matter

AI Engineer with 2+ years shipping production GenAI at Tata Consultancy Services. I embed with problems, build agentic systems and RAG pipelines, and own the unglamorous path to production. MSc Artificial Intelligence at Dublin City University.

View Work → Get in Touch →
Scroll
Key Highlights / 01
0+
Years of experienceshipping production AI & backend systems at Tata Consultancy Services.
0%
Response-time improvementon ServiceNow & Genesys integrations inside GenOps.
0%
Ticket volume reductionfrom a Virtual Agent shipped in production. Automating the repetitive layer.
0%+
F1 on image verificationmulti-modal pipeline with DINOv2, CLIP, SigLIP & calibrated fusion.
About / 02

AI engineer who ships with the customer.

I'm an AI Engineer who works like a forward-deployed builder: sit with the problem, ship GenAI into real workflows, and own what happens after launch. At TCS I built backend systems for GenOps, an internal enterprise AI platform integrating ServiceNow and Genesys, and shipped a Virtual Agent that cut repetitive ticket volume.

Independently I've launched production systems users actually use — multilingual GenAI for Ramakrishna Mission Ireland, multi-agent SVG reasoning, and multi-modal image verification. Outside engineering I run marathons, including the Tata Mumbai Marathon (42.2K) and the Tata Ultra (50K). Distance teaches you to ship.

LocationDublin, Ireland
EducationMSc AI · DCU (2025–)
Emailrahulrajgnr@gmail.com
LinkedInrahulraj1406 ↗
GitHubrahulraj1406 ↗
LanguagesEN · HI · KN
Expertise / 03

A toolkit built for shipping.

Agentic AI
LangGraphLangChain MCPRAG Prompt EngineeringLLM APIs Multi-Agent Systems
ML & Models
PyTorchHugging Face LoRA / PEFTFine-Tuning Computer VisionASR / Speech scikit-learnOpenCV
Languages
PythonTypeScript JavaScriptSQL
Backend & Data
FastAPINode.js ExpressPostgreSQL MySQLMongoDB RedisSSE Streaming
Vectors & RAG
FAISSChroma Pineconepgvector Azure AI Search
Platforms
AzureDocker GitHub ActionsNext.js ReactVercel ServiceNowGenesys
Practice
Customer DeliveryCross-functional Collab Agile ShippingPost-launch Iteration Cost & Latency Tracking
Selected Work / 04

Experience

Jul 2023 to Aug 20252y · Bengaluru
Systems Engineer · GenOps
Tata Consultancy Services
PythonFastAPI Node.jsAzure ServiceNowGenesys
GenOps · Internal AI Platform
  • Built production GenAI applications integrating LLMs into enterprise workflows using Python (FastAPI) and Node.js on GenOps, TCS's internal AI operations platform.
  • Designed and optimised ServiceNow & Genesys API integrations — endpoint configuration and payload filtering — for a 30% improvement in response time.
  • Reduced application processing overhead by 20% through caching and structured backend performance work.
  • Eliminated manual data sync with cron jobs, web-job triggers, and Azure Event Grid event-driven pipelines.
  • Secured APIs with RBAC, OAuth/SSO, token-based auth, and Azure External Identities.
ServiceNow · Conversational AI
  • Shipped a ServiceNow Virtual Agent and Genesys-integrated conversational service desk that cut repetitive ticket volume by 25%.
  • Monitored, debugged, and iterated on deployed AI systems post-launch — prompt tuning and failure-mode analysis with product, data, and operations teams.
  • Built interactive dashboards in Power BI and Angular that reduced manual reporting effort by 80%.
Feb to May 20234 mo · Intern · Bengaluru
Software Engineering Intern
Tata Consultancy Services
PythonNLTK scikit-learnTF-IDF
  • Built an automated NLP scoring and classification pipeline in Python (NLTK, TF-IDF, scikit-learn) over enterprise-scale sustainability / ESG documents.
  • Owned text preprocessing, feature extraction, and model validation workflows end to end.
Aug to Dec 20225 mo · Intern
Machine Learning Intern
Dhilligence Software Services
PythonTensorFlow OpenCV
  • Developed a real-time fire-detection system with Python & TensorFlow, optimising object-recognition accuracy on site.
Selected Projects / 05

Things I've built.

2025 to Present · Production AI Case Study
AI-Powered Image Verification

End-to-end multi-modal pipeline for image forensics and verification. Vision-language models DINOv2, CLIP, SigLIP with LoRA/PEFT domain adaptation, XGBoost on forensic features, and Isotonic Regression calibration. Multi-signal fusion across semantic embeddings, DCT/FFT artefacts, PRNU residuals, and ELA. Served via FastAPI with 93%+ F1 on real evaluation data.

PythonFastAPI DINOv2CLIP SigLIPLoRA XGBoost
2025 · DCU · In preparation Research
Telugu ASR · Low-Resource Speech

Fine-tuned and evaluated five speech architectures (wav2vec2-XLS-R-300M, Whisper-small/medium, MMS-1B-All, Telugu Whisper) on MSCIL Telugu (~44K → 35K cleaned). Out-of-domain eval on FLEURS exposed the in-domain evaluation trap: KenLM gains invert OOD. MMS zero-shot at 40.05% FLEURS WER beat every fine-tuned model. Four months across five compute environments.

PyTorchWhisper wav2vec2MMS KenLMCTC
2025 · Live rkmireland.org ↗
Bhakti Saurabha · Multilingual GenAI

Independently conceived, built, and launched from zero for Ramakrishna Mission Ireland. LLM translation & structuring pipeline: 96 sacred hymns into 4 languages with phonetic transliterations and a structured JSON schema. Full-stack Next.js + Claude API + Vercel. Reached users across 11 countries within days of launch.

Next.jsTypeScript Claude APIPython VercelCI/CD
2025 · Production Multi-Agent GitHub ↗
VectorSight · SVG Spatial Engine

14-step LangGraph multi-agent pipeline turning raw SVG into structured spatial representations for LLM reasoning. RAG over FAISS, MCP-compatible tool server, FastAPI + SSE streaming, and multi-provider routing with per-agent latency, tokens/sec, and cost tracking.

LangGraphMCP FAISSFastAPI SSENext.js
Capstone · 2023 Case Study
Smart Classroom System

BE capstone: real-time sensor-driven classroom automation and computer-vision engagement analytics. Foundation for the IJARIIE 2023 publication on facial-recognition attendance and expression analysis.

PythonML OpenCVAutomation
Education / 06

Education

2025 to Present In progress
MSc in Artificial Intelligence
Dublin City University · Dublin, IE
Machine Learning · Data Analytics · Data Mining · Statistical Analysis · Computational Science
2019 to 2023 Completed
BE, Information Science & Engineering
AMC Engineering College · Bengaluru, IN
Capstone: Smart Classroom System. Python & ML for real-time automation and analytics in education.
Publications / 07

Research & publications.

Manuscript · In preparation2025–26

The In-Domain Evaluation Trap in Low-Resource Telugu ASR

Co-authored with Sai Srikar Devasani; School of Computing, Dublin City University. Fine-tuning and language-model fusion gains that do not transfer out-of-domain. Contribution: ASR fine-tuning, experiments, engineering, methodology, and significant portions of the manuscript.

ASRLow-Resource NLP Whisperwav2vec2 Evaluation
Publication · IJARIIEMay 2023

Evaluation of Student Attendance Using Facial Recognition & Expression Analysis

Designed and implemented a real-time facial-recognition attendance system with engagement analysis via expression recognition. Applied image preprocessing and trained supervised classifiers automating Excel-format attendance reporting.

Facial RecognitionComputer Vision Supervised LearningImage Processing Python
Credentials / 08

Certifications & training.

MCP
Anthropic MCP Certification
Model Context Protocol
↗
AZ
Azure AI Fundamentals (AI-900)
Microsoft Certified
↗
YP
Young Professionals Certification
Tata Consultancy Services
↗
VI
Virtual Internship Program
Accenture
↗
Contact / 09

Let's build something.

Currently pursuing my MSc in AI at Dublin City University and looking for roles as an AI Engineer, ML Engineer, Software Developer, or Systems Engineer — Ireland, remote, or on-site. Happy to talk about shipping GenAI where users depend on it.

rahulrajgnr@gmail.com →
Available immediately Dublin, IE · open to remote · UTC±0
Email rahulrajgnr@gmail.com ↗ LinkedIn linkedin.com/in/rahulraj1406 ↗ GitHub github.com/rahulraj1406 ↗ Phone +353 89 208 0485 ↗ Resume Download PDF ↓