Applied AI Engineer at TCS

I build LLM-powered applications

LLM applications, RAG systems and multi-agent automation in Python. My enterprise AI work has cut manual effort by up to 50%.

Get in touch

Move your cursor over the field: it acts as a search query, and the six nearest vectors light up.

40–50%

less manual effort across enterprise workflows with LLM automation (STC)

30–40%

lower automation cost with a Python + Selenium framework

~45%

less manual effort from XML validation scripts (SBI YONO 2.0)

2

end-to-end Agentic AI PoCs delivered and moved to enterprise VMs

About

I build and deploy LLM-powered applications, RAG systems and AI automation that real teams use every day.

At TCS I work on enterprise AI for telecom and banking clients. I like the full path from idea to production: agents that finish multi-step tasks on their own, retrieval that finds the right context, and automation that removes hours of manual work.

Outside work I compete in hackathons, solve LeetCode problems daily, play chess and practise Arabic and French.

Experience

Systems Engineer, AIML

Tata Consultancy Services (TCS) · Nov 2024 to present

Saudi Telecom Company (STC)

  • Built and deployed LLM-powered automation systems that reduced manual operational effort by 40–50% across enterprise workflows.
  • Designed multi-agent AI pipelines in Python for autonomous task execution and multi-step workflow orchestration.
  • Developed RAG and semantic search systems on a ChromaDB vector database for context-aware information retrieval.
  • Delivered two end-to-end Agentic AI PoCs and migrated them from TCS environments to STC enterprise VMs.
  • Engineered a Python + Selenium automation framework that reduced dependency on UiPath and cut automation costs by 30–40%.
  • Clustered operational emails with K-Means, DBSCAN, TF-IDF and embeddings to surface recurring issues and patterns.
  • Built revenue forecasting models with Prophet and ARIMA for operational planning.

LLMsMulti-agentRAGChromaDBSeleniumK-MeansDBSCANProphetARIMA

State Bank of India (SBI), YONO 2.0

  • Built and maintained RESTful microservices in Spring Boot for a large-scale banking platform.
  • Wrote Python automation scripts for XML validation and workflow automation, reducing manual effort by about 45%.
  • Worked with Android, Flutter and Angular teams to keep API integrations seamless.

Spring BootREST APIsPythonXML

Projects

Agentic AI Investment Research Assistant

  • Built an agentic research platform with LangGraph and LLMs that automates stock analysis, forecasting, risk assessment and report generation.
  • Integrated the Yahoo Finance API and engineered technical indicators for trend analysis and ML-based stock return prediction.
  • Added financial news sentiment analysis and transaction-cost-aware backtesting that compares strategies against buy-and-hold.
  • Built tool-calling workflows with structured outputs, source citations and validation to make the LLM more reliable.

PythonLangGraphLLMsXGBoostScikit-learnPandasyfinanceFastAPIStreamlit

See it on GitHub

Transformer Architecture From Scratch

  • Implemented a Transformer in PyTorch from first principles: token embeddings, positional encoding, multi-head self-attention and feed-forward layers.
  • Assembled the full encoder-decoder architecture and validated forward-pass dimensions.

PyTorchSelf-attentionEncoder-decoder

See it on GitHub

Earlier builds

Skills

GenAI and agents

  • LLM applications
  • LangGraph
  • Multi-agent pipelines
  • Tool calling
  • RAG
  • Semantic search
  • ChromaDB
  • Embeddings
  • LLM evaluation

ML and data

  • PyTorch
  • Scikit-learn
  • XGBoost
  • Pandas
  • K-Means
  • DBSCAN
  • TF-IDF
  • Prophet
  • ARIMA

Backend and automation

  • Python
  • FastAPI
  • Streamlit
  • Spring Boot
  • REST microservices
  • Selenium
  • Java
  • SQL
  • XML

Also built with

  • JavaScript
  • HTML
  • CSS
  • Android
  • Flutter
  • Dart
  • SQLite
  • C
  • GitHub

Honors and achievements

WinnerTCS AI Friday Hackathon

2nd placeCodeathon

QualifiedSmart India Hackathon

1112+ problemsLeetCode, 1043-day streak

Cleared GATE2024 and 2025

All milestonesGoogle Cloud Ready Program

830-day streakLearning Arabic and French on Duolingo

ChessWon several competitions

Certifications

  • Google Cloud: Gen AI Fundamentals
  • Google Cloud: Large Language Models

Leadership and activities

  • STC LLM Evaluation Lead
  • Co-Technical Head, AIMSA
  • App Development Co-Lead, GDSC
  • Mentor, TCS AI Friday Hackathon
  • AI and Automation Floor SPOC
  • Floor Lead, TCS Olympus Building
  • TCS Maitree Spock SPOC

Get in touch

Have a project, a question or an idea in applied AI? Send a message and I'll get back to you.