Skip to content
View shreevershith's full-sized avatar

Block or report shreevershith

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
shreevershith/README.md

Hey, I'm Shreevershith πŸ‘‹

I'm a Software Engineer. I build reliable, scalable systems across backend engineering, automation, and applied AI. Currently open to backend, systems, and applied AI roles.

LinkedIn Portfolio GitHub


🧠 What I Build

  • Backend & Systems β€” APIs, modular architectures, RBAC, workflow logic, performance optimization
  • Applied AI & NLP β€” ML pipelines, RAG, multi-agent orchestration, LLM integration, prompt engineering
  • Automation & Data β€” workflow pipelines, ETL, analytics, Power Platform, decision-support tools
  • Cloud & Observability β€” Azure, AWS, Docker, CI/CD, real-time dashboards, SSE
  • MLOps β€” experiment tracking, model versioning, evaluation frameworks

πŸš€ Featured Projects

πŸ”Ή AkashGuard β€” Open Agents Hackathon 2026 Β· 2nd Place (Akash Track)

Autonomous self-healing agent for decentralized cloud on Akash Network. Monitors services, uses Llama 3.3 70B via AkashML to diagnose failures, and automatically redeploys to new providers with a real-time dashboard and Telegram alerts.
Python FastAPI Akash Network AkashML LLM SQLite Telegram


πŸ”Ή Emotion Detection with Transformers β€” Kaggle 1st Place πŸ₯‡

Multi-label emotion classification with RoBERTa, ALBERT, DistilBERT, and QLoRA fine-tuning on Qwen. 4-bit quantization and threshold optimization under GPU constraints.
Python Transformers QLoRA HuggingFace


πŸ”Ή Portfolio

A responsive developer site showcasing projects, experience, RAG chatbot, dark mode, and linked to Google Analytics.
Next.js React Tailwind CSS Framer Motion


πŸ”Ή DealGraph β€” AWS x Anthropic x Datadog GenAI Hackathon 2026

AI-powered due diligence copilot for VC investors. Upload a pitch deck and get confidence-scored analysis with verified claims, competitive intelligence, and a voice-narrated deal memo. Claim-level routing to graph, web, and LLM verification.
Python FastAPI Strands Agents Next.js Memgraph LLM D3.js


More on GitHub β€” GlassesToReels, Cactus Voice, LegacyCouncil, and others.


πŸ›  Tech Stack

Languages
Python JavaScript C# SQL

Backend & Frontend
FastAPI .NET REST Next.js

AI / ML
PyTorch scikit-learn HuggingFace MLflow Azure AI

Cloud, Data & DevOps
Azure AWS Docker MongoDB PostgreSQL


πŸ“œ Certifications

Cert Issuer
Azure Fundamentals (AZ-900) Microsoft
Azure AI Fundamentals (AI-900) Microsoft
Azure Data Fundamentals (DP-900) Microsoft
Power Platform Fundamentals (PL-900) Microsoft
AWS Certified Developer (DVA-C02) AWS (in progress)

🀝 Let's Connect

I'm open to roles in backend engineering, systems engineering, and applied AI, especially where clean abstractions and real-world reliability are important.

πŸ“¬ LinkedIn Β· 🌐 Portfolio

Pinned Loading

  1. nutriguard-api nutriguard-api Public

    NutriGuard is a guarded nutrition estimation API (FastAPI) that combines deterministic food data with a hybrid RAG architecture to provide explainable calorie estimates while minimizing LLM usage a…

    Python

  2. emotion_detection_using_transformers_kaggle emotion_detection_using_transformers_kaggle Public

    Emotion detection experiments using Transformers (RoBERTa, ALBERT, DistilBERT) and QLoRA-based fine-tuning on Qwen models. Features multi-label classification, 4-bit quantization, and threshold opt…

    Jupyter Notebook

  3. 2DTableTennisGameUsingGLUT 2DTableTennisGameUsingGLUT Public

    A 2D animation based on the table tennis game using C++ and glut package with the help of Visual Studio/CodeBlocks software.

    C++ 1

  4. EstimationOfPhotometricRedshiftUsingDL EstimationOfPhotometricRedshiftUsingDL Public

    The objective of the project is to find the Photometric Redshift from the data extracted from SDSS server. Data is extracted from server through SQL. The extracted data is sampled with many machine…

    Jupyter Notebook 1

  5. CustomerAnalyticsUsingMLE CustomerAnalyticsUsingMLE Public

    Constructed Poisson and NBD models to analyze billboard exposures

    Jupyter Notebook