engineer = {
"name": "Dodda Jashwanth Sai",
"role": "AI Engineer",
"focus": ["LLMs", "AI Agents", "RAG", "Machine Learning", "Deep Learning"],
"foundations": ["Data Science", "Data Structures & Algorithms", "System Design"],
"builds": ["ML/DL models", "AI workflows", "autonomous agents", "self-hosted AI systems"],
"open_source": ["MoCICD", "CanIRunLLM"],
}AI Engineer with strong fundamentals in machine learning, deep learning, and data science. I build models, LLM-powered workflows, and agents, and ship them as reliable, production-ready services, including self-hosted deployments. My work spans healthcare and education, and I publish open-source tools for the AI ecosystem.
Each area below shows what I build and the skills behind it.
I can build
- Classification, regression, and prediction models on real datasets
- Neural networks trained, tuned, and evaluated end to end
- Computer vision pipelines for image and video tasks
- Data cleaning, feature engineering, and analysis workflows
Skills & tools
Neural Networks Computer Vision NLP Feature Engineering Model Evaluation Data Analysis
I can build
- LLM-powered applications with structured, reliable outputs
- Fine-tuned and optimized models for specific tasks
- Self-hosted LLM services with OpenAI-compatible APIs and SDKs
- Prompt pipelines with function calling and guardrails
Skills & tools
Prompt Engineering Function Calling Structured Outputs Fine-Tuning (LoRA) Quantization Model Serving
I can build
- Stateful multi-agent workflows that plan, call tools, and hand off tasks
- Role-aware agents with memory and personalized behavior
- Automation for evaluation, compliance, and document-heavy processes
Skills & tools
Agentic Workflows Tool Use Function Calling Contextual Memory Guardrails Role-Based Access
I can build
- Document ingestion pipelines for PDF, DOCX, and text
- Semantic search and question answering over private knowledge
- Knowledge bases that stay continuously in sync with their sources
Skills & tools
Embeddings Semantic Search Document Chunking Information Retrieval Named Entity Recognition Zero/Few-Shot Classification
I can build
- Systems that understand text, speech, and video together
- Video-to-video retrieval, answering questions with contextual video responses
Multimodal LLMs Speech Extraction Video Processing Multimodal Retrieval
I can build
- Async, low-latency APIs that put models and agents in front of users
- Secure, multi-tenant services with authentication and role-based data isolation
- Containerized, cloud-deployed AI services with CI/CD
Skills & tools
REST APIs Microservices Async Processing OAuth RBAC System Design
Data Structures & Algorithms · Object-Oriented Programming · DBMS · Operating Systems · Computer Networks · System Design
|
MoCICD · Continuous model discovery, sandboxed benchmarking, and human-approved promotion for every place an application calls an LLM.
|
CanIRunLLM · Scans your hardware and tells you which open-source LLMs your machine can run, with a local dashboard and one-click model setup.
Repository · PyPI · Live site |
| Project | What it does | Built with |
|---|---|---|
| Self-Hosted LLM Platform | OpenAI-compatible model API with an internal SDK and API-key access, so apps can move off third-party providers | vLLM FastAPI Quantization |
| Live Sync RAG Platform | Keeps enterprise knowledge bases continuously synced and searchable over PDF, DOCX, and text | RAG Vector DB OAuth Async Python |
| Video-to-Video RAG | Answers questions over training videos by generating contextual video responses | Multimodal Retrieval Embeddings LLMs |
| AI Education Platform | Multi-tenant platform with stateful multi-agent workflows and role-based data isolation | LangGraph Function Calling RBAC |
| AI Telehealth Platform | LLM agents for clinical documentation and entity extraction with guardrails, connected to EHR databases | LLMs NER Async APIs |
- B.Tech, Computer Science (AI & ML), Marri Laxman Reddy Institute of Technology and Management
- Certifications: Google Cybersecurity, Machine Learning, Artificial Intelligence, Data Science & Analytics
- Selected for the National Innovation Design and Entrepreneurship Bootcamp 2024
- 300+ coding problems solved across LeetCode, HackerRank, and CodeChef
- Conducted AI/ML workshops, hackathons, and mentoring sessions for students and junior developers