HELLO_WORLD.INIT()

SAIKRISHNA

A COMPUTER SCIENCE ENGINEER OBSESSED WITH BUILDING SCALABLE SYSTEMS AND BREAKING DESIGN NORMS.

VIEW_PROJECTSCONTACT_ME
/// FULL STACK DEVELOPER /// AI ENGINEER /// SYSTEM ARCHITECT /// CREATIVE CODER /// /// FULL STACK DEVELOPER /// AI ENGINEER /// SYSTEM ARCHITECT /// CREATIVE CODER /// /// FULL STACK DEVELOPER /// AI ENGINEER /// SYSTEM ARCHITECT /// CREATIVE CODER /// /// FULL STACK DEVELOPER /// AI ENGINEER /// SYSTEM ARCHITECT /// CREATIVE CODER ///

About Me

I am a pre-final year Computer Science and Engineering student at Vellore Institute of Technology, Chennai. My technical journey is driven by a passion for building scalable backend systems and exploring the potential of Artificial Intelligence.

Currently serving as a Tech Team Member at ACM-VIT Chennai, I have hands-on experience in delivering REST APIs, optimizing deployment workflows, and leading agile engineering teams. I thrive in environments that challenge me to optimize latency, ensure reliability, and design intuitive developer experiences.

When I'm not coding hackathon-winning projects or engineering deepfake detectors, I'm mentoring peers in API design and Git workflows.

System_Log

2024 - NOW

Tech Team Member

ACM-VIT
  • Delivered 5+ REST APIs and internal tools.
  • Engineered backend systems in C/Python.
  • Reduced onboarding time by 60% via doc restructuring.

Project_Archive_

202501

AI Website Builder

AWS LambdaBedrockNext.js
  • Natural-language → React Code Pipeline.
  • Automated CI/CD via S3 + CloudFront.
202502

Deepfake Detector

PyTorchCNNPython
  • 98.5% Accuracy on Video Streams.
  • Real-time processing < 200ms.
202503

CosmoMap AI

YOLOv8FastAPILeaflet
  • Aerial Object Detection.
  • Geospatial Visulization.
202504

Neural Global Illumination Engine

PyTorchMLPRay TracingCUDA
  • Reframed Global Illumination as a regression task — MLP predicts radiance instead of ray-tracing.
  • Brought per-frame compute from 26ms to 11ms with +4.1dB PSNR improvement. 60+ FPS with dynamic lighting.
202505

PowerBench-Consumer

PyTorchTRLGRPOLLaMA
  • Benchmarked LLM inference and GRPO training on RTX 2050 (4GB VRAM), extending DREAM:Lab's Jetson Orin AGX methodology.
  • Key finding: INT8 inference is ~3x slower than FP16 on consumer GPUs — quantization gains are not portable across architectures.
202506

NEXUS — Neural Embedding Search

PythonHNSWSimHashFastAPI
  • Semantic search with soft cluster assignments, entropy-adaptive cache thresholds, and two-level caching.
  • In-process HNSW for sub-linear lookup with full research-style write-up on trade-offs and failure modes.

Tech_Stack_Db

Languages

JavaScriptPythonCC++GolangSQL

Frameworks

ReactNode.jsExpressDjangoFastAPI

Databases

MongoDBMySQLPostgreSQL

AI/ML

PyTorchNumPyPandasOpenCV

DevOps/MLOps

GitDockerLinuxMLflowDVCAWSAzure

Achievements

🏆

Won MongoDB Track & 4th Overall

DevsHouse25

🥈

Finalist

Qualcomm Edge AI Hackathon 2025

Education

B.Tech in Computer Science and Engineering

2023 – 2027 (Expected)

Vellore Institute of Technology, Chennai

Get In Touch

I'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions.

EmailLinkedInGitHub