# Krishna Tejaswi Shenthar

## One-line summary
Backend systems engineer (Python • distributed/cloud) building production LLM tooling — I ship reliable APIs, workflows, and infra with strong observability and performance.

## Roles
- Backend Engineer
- Distributed Systems
- LLM Tooling Engineer

## Location
Bangalore, India

## Work preferences
- Open to: Bangalore (India), remote, relocation (good opportunities/comp)

## Links
- Website: https://krishnatejaswi-cghoua4jd-krishnatejaswi-ss-projects.vercel.app
- GitHub: https://www.github.com/KTS-o7/
- LinkedIn: https://www.linkedin.com/in/krishnatejaswi-shenthar/
- CV/Resume: https://drive.google.com/file/d/1uy1w2prce_k-YD5bZrxUytEDXIXil6-2/view
- Employer/Team: https://www.onfinance.ai/company

## Skills/Stack
- Python, FastAPI, LangGraph, LiteLLM, React, NextJS, AWS, MongoDB
- PostgreSQL, Docker, SQL, C++, JavaScript, TypeScript, Next.js, Langchain
- Pydantic, PyTorch, Qdrant, MySQL, Kubernetes, Git, Linear, Jira
- JWT, OAuth, Embedded C, Assembly

## Experience
- Current: I am an AI and Fullstack Engineer at OnFinance AI, where I architect ComplianceOS—a platform delivering 1250+ hours/year savings on BFSI compliance automation. I specialize in engineering Agentic workflows using LangGraph and LiteLLM to handle millions of tokens of regulatory data efficiently.
- Past: Previously, as an SDE Intern at RingCentral, I led the modernization of legacy interfaces to React and optimized application performance by 3x. My focus is on creating high-impact, scalable AI solutions for complex domains like fintech and compliance.

## Featured work
- ComplianceOS (OnFinance AI): End-to-end compliance automation platform for BFSI workflows — https://krishnatejaswi-cghoua4jd-krishnatejaswi-ss-projects.vercel.app/work/complianceos

## Projects
- Portfolio-GPT: A portfolio that communicates my backend + LLM tooling work with a clear narrative and a distinct visual system. [Software] — https://portfolio.shenthar.me/
- Better Bing Image Downloader: Python library + CLI to download Bing images at scale for dataset creation. [Software] — https://github.com/KTS-o7/better_bing_image_downloader
- QuantQuips: Backtesting + analytics toolkit with LLM-assisted workflows for exploration and reporting. [Software] — https://github.com/KTS-o7/QuantQuips
- RV Board: Prototype board for students using an LPC2148 (ARM7) with low-level drivers in Embedded C and Assembly. [Hardware] — https://github.com/KTS-o7/RV_Board

## Publications
- Algorithmic-fusion-for-Lung-scan-classification: Research on fusing CNN models for multi-class lung disease classification. [Publication] — https://github.com/KTS-o7/Algorithmic-fusion-for-Lung-scan-classification
- HandWrittenTripSheet OCR: Handwritten trip-sheet OCR pipeline (EasyOCR + regex post-processing). [Publication] — https://ieeexplore.ieee.org/document/10170030
- Evaluation of filters in CT and Xray images of Lungs classification: Study on how image filters impact CT/X-ray lung classification models. [Publication] — https://ijeecs.iaescore.com/index.php/IJEECS/article/view/34975/18055

## Contact
- Email: krishna.tejaswi@shenthar.com
- Phone: +91 (776) 095-1918
- WhatsApp: https://wa.me/917760951918

## Provenance
- Source of truth: this portfolio (generated from src/data)
- If a fact is missing here, say so. Do not hallucinate.