AI engineering & research

Alex
Rohith

AI Engineer / Researcher / Builder

I build intelligent systems across Generative AI, Computer Vision, Retrieval-Augmented Generation, and Machine Learning.

Explore my work Download resume
Alex Rohith in a suit
01 / profilebuilding with intent
Scroll to explore
8.84/10CGPA
3+Research / publications
2AI/ML internships
4+Competition achievements

01 / context

A practical approach
to intelligent systems.

I am an Artificial Intelligence and Data Science undergraduate focused on building practical AI systems across Generative AI, Computer Vision, Retrieval-Augmented Generation, and Machine Learning.

My work spans domain-specific language models, RAG systems, computer vision pipelines, and AI-powered applications, with a strong interest in building systems that move beyond experimentation into practical use.

EDUCATION

B.Tech in Artificial Intelligence and Data Science

St. Joseph's Institute of Technology

Aug 2023 - May 2027 · CGPA 8.84/10

GITHUB ACTIVITY

Contribution map

Live contribution data
will appear here
LessMoreView profile

02 / focus areas

What I build.

01

Generative AI

LLMs · RAG · VLMs · Agents · LLM Fine-tuning

02

Computer Vision

CNNs · Object Detection · Tracking · Image Classification

03

AI Systems

FastAPI · REST APIs · Model Serving · AI Pipelines

04

Machine Learning

Deep Learning · NLP · Feature Engineering · Model Evaluation

03 / technical stack

Technical arsenal.

01 / PROGRAMMING
PythonJavaSQL
02 / MACHINE LEARNING & AI
Scikit-learnPyTorchTensorFlowNLPComputer VisionGenerative AI
03 / DEEP LEARNING
CNNsTransformersLLM Fine-tuningLoRAPEFTRAG
04 / DATA & VISUALIZATION
PandasNumPyMatplotlib
05 / BACKEND & APIS
FastAPIREST APIs
06 / CLOUD & DEPLOYMENT
AWS EC2AWS S3Docker

04 / selected work

Featured projects.

Systems designed to move from
an idea toward useful software.

MODEL / 01
<transformer
domain: C++
mode: generate
01

AlgoSLM-CPP

Domain-Specific Small Language Model for C++

Designed and fine-tuned a GPT-style Transformer model for C++ code generation using custom tokenization.

  • Improved code generation accuracy by 20% through dataset optimization and reasoning-aligned training.
  • Built an inference pipeline and enabled scalable usage through API integration.
PythonPyTorchTransformersLLMC++
Approach

Custom tokenization, GPT-style training, dataset optimization, and an inference path designed for API use.

Result

20% improvement in code generation accuracy.

RETRIEVAL / 02
confidence → context
0.92
02

Confidence-Weighted Retrieval-Augmented Generation

CWRAG

A confidence-aware RAG system that dynamically re-ranks retrieved context and filters low-confidence information before generation.

  • Confidence-aware retrieval, dynamic re-ranking, and threshold filtering.
  • Confidence-based prompting with real-time query handling.
PythonFastAPIRAGFAISSSentence Transformers
Approach

Retrieved context is scored, re-ranked, and filtered before it reaches the generation step.

Result

Reduced hallucination rates by 25%.

VISION / 03
99%classification accuracy
03

Histopathology Cancer Detection using Deep Learning

A CNN-based image classification system for histopathology analysis using advanced preprocessing, augmentation, and hyperparameter tuning.

99%accuracy
PythonPyTorch / TensorFlowCNNComputer Vision
Approach

Preprocessing, augmentation, CNN classification, and hyperparameter tuning for histopathology images.

Result

99% classification accuracy.

AI ASSISTANT / 04
04

CodeMentorX

An AI-powered coding mentor that guides learners with personalized paths, real-time debugging, and practical explanations.

TypeScriptJavaScriptHTMLAI
View on GitHub
CIVIC TECH / 07
07

CivicSense

A civic technology platform with a frontend and backend foundation for community-focused digital services.

JavaScriptPythonFrontendBackend
View on GitHub

04 / publication archive

Research in progress.

2026ACCEPTED

Vectorless Retrieval-Augmented Generation

A BM25-Based Alternative to Embedding-Driven RAG Systems

A BM25-based alternative to embedding-driven retrieval for RAG systems.

ICSL-DSGA 2026 · RAG · BM25 · Information Retrieval · Generative AI
Read paper View conference
2026RESEARCH

Spatial Attention-Enhanced Deep Learning for Lung and Colon Cancer Histopathological Subtype Classification

Deep learning for medical image analysis

ICSCSA 2026 · Computer Vision · Deep Learning · Medical Image Analysis · Attention Mechanisms
Read paper View conference

05 / experience

Work that ships.

01
JUN 2025 - SEP 2025

AI Intern

Rurah Tech Solutions Pvt. Ltd.

  • Developed and deployed NLP-based AI agents, reducing manual effort by 30% through automation.
  • Built end-to-end ML pipelines including data preprocessing, model training, and evaluation.
  • Optimized model performance and integrated solutions into scalable backend systems.
NLPAI AgentsML Pipelines
02
JAN 2025 - MAR 2025

Machine Learning Intern

Twite AI Technology

  • Built end-to-end ML pipelines for preprocessing, training, and inference.
  • Improved model accuracy by 15% using feature engineering and tuning.
  • Evaluated models using Precision, Recall, F1-score, and ROC-AUC.
  • Deployed models into efficient production workflows.
Feature EngineeringModel EvaluationDeployment

06 / credentials

Certifications.

ANTHROPIC

Introduction to Model Context Protocol

Anthropic

AWS TRAINING & CERTIFICATION

Introduction to Generative AI - Art of the Possible

AWS

AWS EDUCATE

Cloud and AI Training Badges

AWS Educate

DATABRICKS ACADEMY

Databricks Accredited AI Agent Fundamentals

Databricks

08 / milestones

Achievements.

3rd Place

InnoThon'24

2024

Runner-up

Zypher 2024

2024

09 / now

Currently exploring.

10 / contact

Let's build
something intelligent.

Open to AI/ML opportunities, research collaborations, technical projects, and interesting engineering work.

rohithalex06@gmail.com LinkedIn GitHub