S.Ajay Resume
Available for opportunities

Somala Ajay

AI/ML Engineer with 3+ years of experience. Currently at TCS designing ML pipelines. Delivered 20+ AI/ML projects across NLP, Computer Vision, Deep Learning, GenAI & RAG pipelines.

Get in Touch Resume ↓ LinkedIn GitHub
20+
AI/ML Projects
87%
Model Accuracy
6+
Certifications
3yr
Experience
About Me

Who I am

Somala Ajay

AI/ML Engineer · Python Developer · Data Scientist

AI/ML Engineer with 3+ years of experience building ML pipelines, NLP systems, and Generative AI applications. Currently at TCS as Associate Software Engineer, designing and deploying production ML models using Scikit-learn, Flask REST API, and SQL. Delivered 20+ end-to-end AI/ML projects across NLP, Computer Vision, Deep Learning, and GenAI.

Current Status

Currently at TCS (serving notice). Actively seeking AI/ML Engineer & GenAI Engineer roles in Hyderabad or remote.

20+AI/ML Projects
87%Model Accuracy
3yrExperience

Location

Hyderabad, India · 17.3850° N · IST

Core Stack

Python · TensorFlow · FastAPI · Flask · RAG · FAISS · LangChain · Docker · Scikit-learn · SQL

Education

B.Tech Electronics & Communication Engineering — Aditya University, Surampalem (2018–2022)

candidate_match.py
Work History

Experience

Dec 2025 — PresentCurrent · Notice Period

Associate Software Engineer

Tata Consultancy Services (TCS) — Hyderabad

ML Pipelines · Flask REST API · EDA · SQL · Scikit-learn · CTC: 6 LPA

  • Designed and implemented end-to-end ML pipelines for structured client datasets covering data ingestion, preprocessing, and model training.
  • Built classification models using Scikit-learn, achieving 82–87% accuracy on validation datasets.
  • Deployed ML model via Flask REST API for internal inference workflows.
  • Automated data preprocessing pipeline, reducing manual effort by ~25–30%.
  • Performed EDA, feature engineering using Pandas and NumPy on datasets of 10K–50K records; used SQL for data extraction and transformation.
Sep 2024 — Nov 2025Full-time

Associate

Tech Mahindra — Hyderabad

Enterprise Data & WFMS · Data Integrity · Data Validation · Doc Audit · Data Cleaning

  • Worked on the client project using the Workflow Management System (WFMS) to handle and process structured data.
  • Performed doc auditing and data correction to maintain high accuracy and data integrity.
  • Supported data validation and quality checks, reducing discrepancies and improving workflow reliability.
  • Coordinated with team to manage high-volume data operations efficiently, contributing to streamlined process execution.
Aug 2022 — Jun 2024Full-time

AI/ML Data Scientist

BO IT Solutions Private Limited — Hyderabad

End-to-End AI/ML Development · NLP · Computer Vision · Deep Learning · GenAI

  • Delivered 20+ end-to-end AI/ML projects covering NLP, Computer Vision, and Deep Learning — full pipeline from data collection to model deployment.
  • Performed EDA and feature engineering on datasets of 10K–100K+ records using Pandas, NumPy, and Matplotlib, reducing data preprocessing time by ~30%.
  • Built and evaluated classification and regression models using Scikit-learn and TensorFlow, achieving 80–90% accuracy across multiple project deliveries.
  • Implemented CNN and LSTM-based deep learning models for image classification and sequence prediction, improving baseline accuracy by ~15% over traditional ML.
  • OJT as Assistant Trainer at Vikas Engineering College — demonstrated 20+ projects to students and faculty.

Validation accuracy — TCS classification model

Epoch 161%
Epoch 1074%
Epoch 2082%
Epoch 30 (final)87%

How the pipeline actually runs — TCS ML pipeline

IngestPreprocessTrainDeploy

Illustrative — mirrors the ingestion → preprocessing → training → Flask API deployment structure described above

Technical Stack

What I work with

AI / ML / Deep Learning

Scikit-learn90%
TensorFlow / Keras85%
CNN / LSTM82%
NLP88%

GenAI & RAG

RAG / FAISS88%
LangChain85%
LLMs / Prompt Eng.80%
AI Agents78%

Backend & Data

Python90%
FastAPI / Flask86%
Pandas / NumPy84%
SQL80%

Tools & Infra

Docker82%
Git / GitHub88%
MySQL / SQLite80%
Streamlit75%

Live demo

Try a toy sentiment classifier

A lightweight keyword-scored demo (not a production model) — type a sentence and see it score positive / neutral / negative in your browser.

Positive
0%
Neutral
0%
Negative
0%

Runs entirely client-side · no data leaves your browser

Portfolio

Things I've built

01

PinGuru — Instagram DM Automation SaaS

Python · FastAPI · MongoDB · Instagram Graph API · DigitalOcean · Stripe · JWT

Full-stack Instagram automation SaaS (pinguru.me) with real-time DM triggers, multi-tier Stripe billing, FastAPI/MongoDB backend, JWT/OTP security, and automated comment-to-DM workflows.

View on GitHub →
02

NASA Manual QA System

Python · FastAPI · FAISS · Llama 3.3 · Groq · RAG · sentence-transformers

Production RAG pipeline querying the 270-page NASA Systems Engineering Handbook. 885 semantic vectors, confidence scoring (0–100%), multi-turn history, section-level citations, NASA acronym expansion. i2e Hireathon 2026.

View on GitHub →
03

Brain Tumor Detection

Python · TensorFlow · OpenCV · CNN

CNN-based MRI classifier achieving 90%+ accuracy — detects and classifies brain tumors using deep learning and advanced image processing.

04

Image Caption Generator

Python · CNN · LSTM · FLICKR Dataset · NLP

End-to-end image captioning system combining CNN feature extraction with LSTM sequence generation for automatic descriptive caption generation.

05

Amazon Recommendation System

Python · Pandas · Scikit-learn · Collaborative Filtering

Product recommendation engine using collaborative filtering and content-based methods on Amazon dataset to surface personalized suggestions.

06+

Full Portfolio — 20+ Projects

Python · Flask · Docker · TensorFlow · GenAI · NLP · CV

Stock Prediction, Face Mask Detection, Sentiment Analysis, Food Nutrition Adviser, Virtual Mouse & more — all on GitHub.

View all →
Credentials

Certifications

AI Data ScientistNASSCOM
Career Essentials in Generative AIMicrosoft & LinkedIn
Generative AI IntroductionMicrosoft Learning
Develop an AI App with Azure AI Foundry SDKMicrosoft
Python (Basic)HackerRank
Agile Methodology Virtual ExperienceForage
Learning Microsoft 365 Copilot for WorkLinkedIn

i2e Hireathon 2026 — National AI Hackathon

i2e Consulting · Bangalore · Complex Technical Manual QA System. Production-grade RAG for the NASA Systems Engineering Handbook — FAISS, LLaMA 3.3-70B via Groq, confidence scoring, multi-turn history, section-level citations.

View on GitHub →

B.Tech — Electronics & Communication Engineering

Aditya University, Surampalem, Andhra Pradesh · August 2018 – June 2022

Capstone: Quality Analysis of Rice Granules using MATLAB Image Processing

Get In Touch

Let's connect

Open to new opportunities, collaborations, or just a good tech conversation. Seeking AI/ML Engineer & GenAI Engineer roles.

Hyderabad, Telangana 500081