Hi, I'm
Varshith.
B.Tech Data Science student and full-stack builder. Founder of OQENS and CTO at Unifesto, obsessed with shipping production-grade platforms and AI tools.
Fuelled by open source, cloud native architectures, and raw data. Expertise in React.js, Node.js, PostgreSQL, GCP, and OCI.
About
B.Tech Data Science student building full-stack web applications and AI-powered tools.
Founder of OQENS — an independent search engine ecosystem. Currently CTO at Unifesto, leading technology strategy and cloud infrastructure across GCP and OCI.
Self-driven builder who leverages React, Node.js, and PostgreSQL to ship real, production-grade products from scratch.
Experience
Chief Technology Officer
Founder & Lead Developer
Founder
Core Technologies
AI Infrastructure
Engineered scalable AI infrastructure using highly-optimized, self-hosted LLM inference systems to deliver sub-second reasoning responses.
Deployed Agent Network
Handles complex multi-turn logic, orchestrates tool calling, and generates final user responses. Quantized to GGUF Q5_K_M for optimal VRAM usage.
Specialized agent reserved for high-accuracy algorithmic tasks and structured JSON outputs. Runs on Q4_K_M for rapid inference.
Lightweight model used exclusively for rapid prompt classification, safety filtering, and RAG query generation before hitting heavier models.
Backend Inference Engine
llama.cpp
Highly optimized C++ inference backend maximizing VRAM efficiency on edge hardware.
Ollama
Streamlined local runner for quantized GGUF models, achieving sub-second Time-To-First-Token.
FastAPI Orchestration
High-throughput asynchronous Python endpoints handling concurrent LLM streams, request queuing, and dynamic load balancing across multiple inference nodes. Built to scale proprietary RAG pipelines.
Cloud & DevOps
Highly available
deployments.
Architecting serverless ecosystems, managing high-availability EC2 clusters, and provisioning scalable RDS databases tailored for production workloads.
Deploying intelligent compute engines, orchestrating containerized workloads via GKE, and leveraging native GCP machine learning infrastructure.
Integrating enterprise-grade App Services, configuring secure active directories, and managing resilient cloud storage and CDN solutions.
Setting up high-performance bare-metal servers, automated webhook deployment pipelines, and robust virtual cloud networks for extreme scalability.
Writing infrastructure as code, constructing multi-stage CI/CD GitHub Actions pipelines, and configuring hardcore Systemd & Nginx reverse proxies.
Featured Work
ORENAI
Engineered scalable AI infrastructure using Groq and Llama-3 to deliver sub-second reasoning responses. Orchestrated custom RAG pipelines and intelligent tool execution.
Unifesto
Architected a high-throughput multi-tenant platform for university fests. Designed fault-tolerant PostgreSQL schemas with Prisma and ultra-fast Next.js caching layers.
OQENS
Built a visually stunning, performant agency portfolio prioritizing extreme Lighthouse scores. Integrated advanced Framer Motion layout animations and React-Three-Fiber models.