Sai Pavan
AI/ML Engineer
Overview
Social Links
About
Hello! I am Sai Pavan — a Energetic and innovative Generative AI Engineer with 2 years of experience specializing in AI/ML frameworks, MLOps, containerization (Docker), and orchestration using Kubernetes and Cloud platforms (AWS, GCP). Aims to apply technical skills and creativity in developing state-of-the-art AI solutions within a dynamic team setting.
Let's connect and collaborate!
Tech Stack
Python
MCPs
Git
Docker
MySQL
MongoDB
Redis
ChatGPT
C++
AWS
Google Cloud Platform
Langgraph
Langfuse
LangChain
LlamaIndex
Flask
FastAPI
Socket.IO
WebRTC
Gradio
Streamlit
Selenium
ComfyUI
TensorFlow
PyTorch
Pipecat
Daily
ChromaDB
Firestore
Qdrant
Weaviate
Pinecone
OpenSearch
Neo4j
NebulaGraph
Kubernetes
AWS Bedrock
AWS SageMaker
AWS Lambda
GCP Vertex AI
Hugging Face
JavaScript
Experience
Sify Technologies Limited
Current Employer-
Developed and deployed AI/ML solutions, focusing on data processing, predictive analytics, and automation.
-
Collaborated with cross-functional teams to integrate AI models into scalable enterprise-level solutions, improving operational efficiency and decision-making.
-
Conducted advanced research and prototyping to integrate and scale AI solutions into learning management systems and standalone applications.
-
Customized and deployed open-source AI models on private cloud infrastructure for scalable inference, utilizing custom scripts, XInference, Ollama, Nvidia/TensorRT-LLM and vLLM.
- Python
- Docker
- Kubernetes
- AWS
- GCP
- Dify
- RAG
- Gen AI
- Socket IO
- WebRTC
- Streamlit
- Selenium
- ComfyUI
- TensorFlow
- PyTorch
- Pipecat
- Daily
- Open-source and proprietary LLMs
- Multi-models
- LlamaIndex
- Gradio
- Flask
- FastAPI
- Langchain
- Langgraph
- Langfuse
-
Gained hands-on experience with Generative AI technologies, tools, and frameworks, including model inference and deployment.
-
Developed hands-on proficiency with AWS cloud and subsequently cleared the AWS Certified Developer – Associate certification through practical learning.
- GenAI
- LLMs
- AI frameworks
- AWS
- AWS Certified Developer – Associate
Cisco Networking Academy
-
During this internship, I worked on understanding the fundamentals of computer networking and IP services, including network design, IP addressing, and routing protocols. The primary focus was to help develop an in-depth knowledge of Cisco’s Routing and Switching technologies, which are vital for building scalable and secure networks.
-
Networking & Routing Configuration: Gained hands-on experience in configuring and troubleshooting IP routing protocols, while developing a solid understanding of networking models such as the OSI and TCP/IP stacks, and subnetting for efficient IP address management.
-
Switching & Network Services: Configured VLANs, and Ethernet switching to ensure optimal network performance, and implemented essential IP services like DHCP, DNS, and NAT to support dynamic address allocation, domain resolution, and secure communication across the network.
- Networking
- Routing
- Switching
- CCNA
Educations
- Bachelor of Engineering in Computer Science; CGPA: (8.70/10.0)
- Diploma in Computer Engineering; Percentage: (89.69/100)
- Board of Secondary Education; CGPA: (9.7/10.0)
Projects(5)
- Search Engine - Architected a cloud-agnostic solution combining open-source and commercial components, leveraging vector databases and GCP Vertex AI services. Also includes an Elasticsearch sandbox POC.
- Implemented data enrichment pipelines for metadata tagging, data refinement, spell check, and missing-text interpolation to enhance recall and content quality across text and image assets. Asset Captioning & Descriptions
- Developed custom celebrity recognition using Histogram of Oriented Gradients (HOG) features automatically identify and tags individuals in new images, with support for adding new celebrities. Enabling retraining with curated labels and thresholded inference for precision.
- GCP
- Vertex AI
- NLP
- Histogram of Oriented Gradients (HOG)
- Google Vision API
- Elasticsearch
- Firestore
- Embedding Models
- Vector Databases
- Open-source and proprietary LLMs
- Knowledge Graphs
- Developed an advanced Conversational AI system with a virtual avatar capable of mimicking human voice, expressions, and emotions, delivering highly interactive, personalized, and context-aware user experiences. Integrated Retrieval-Augmented Generation (RAG) and search capabilities to provide dynamic, immersive, and emotion-driven interactions across various scenarios with both visual and verbal feeds.
- Built in three versions: Unity WebGL for integration with SkillFLO, Unreal Engine for immersive Experience Center demo, and Unity EXE for DevLearn Conference presentations.
- Demos: NORA (Next-gen Operational Response Agent) presented at DevLearn, and PUJA (Personal Universal Job Assistant) designed for Experience Center as scenario-driven avatar.
- Histogram of Oriented Gradients (HOG)
- Retrieval-Augmented Generation (RAG)
- Gen AI
- Real-Time Communication (RTC)
- Socket IO
- Python
- FastAPI
- Integration and deployment of Gen AI models
- Content Automation pipelines
- Backend Development
- Architected an agentic AI workflow using LangGraph (Python) to automate ticket resolution within Sify’s cloud management ecosystem, introducing intelligent state-based decision-making.
- Developed a proof-of-concept for automated infrastructure provisioning on AWS using MCPs provided by AWS, significantly streamlining and accelerating ticket handling workflows.
- Creating MCPs like an SSH server and a code generator (generator, debugger, executor) to streamline communication between the LLM and Sify’s cloud infrastructure.
- Building a fully autonomous ticket management system leveraging agentic states and dynamic nodes to process, diagnose, and resolve incidents — escalating to human intervention only when necessary.
- Implemented multiple MCP servers to enable seamless communication between AI agents and Sify’s cloud infrastructure, ensuring reliable orchestration and data exchange across environments.
- Architected a serverless generative AI platform for e-learning using LlamaIndex, LangChain, and AWS, enabling semantic search, automated assessment generation, storyboard creation, and AI-enhanced media libraries.
- Developed and integrated AI-driven modules for automated content creation (HTML, video, storyboards) and assessment generation from multiple sources using NLP techniques.
- Engineered AI-powered analytics for custom report generation via text prompts with visual insights, and implemented an AI chatbot to enhance content discoverability and user access to learning resources.
- Python
- REST APIs
- Gen AI Models Integration and deployment
- Content Automation pipelines
- Backend Development
- Prompt Engineering
- Developed an AI-powered interactive assistant for children (ages 4-12) using GPT-4o, Python, and FastRTC, delivering real-time bedtime stories, homework help, and engaging conversations.
- Built a safe, age-appropriate interface with dynamic multi-agent context switching and adaptive storytelling for personalized learning experiences.
- Introduced imaginative roleplay and AI-generated quizzes, fostering creativity, engagement, and knowledge retention through immersive interactions.
- GPT-4o
- Python
- Multi-Agent Systems
- Real-time AI
- WebSockets
- LSTM (Long Short-Term Memory)
Honors & Awards(2)
Certifications(7)
AWS Certified AI Practitioner
- Issued by
- Amazon Web Services (AWS)
- Issued on
GCP Certified Generative AI Leader
- Issued by
- Google Cloud Platform (GCP)
- Issued on
AWS Certified Developer - Associate
- Issued by
- Amazon Web Services (AWS)
- Issued on
Oracle Cloud Infrastructure 2025 Certified AI Foundations Associate
- Issued by
- Oracle
- Issued on
Oracle Cloud Infrastructure 2025 Certified Generative AI Professional
- Issued by
- Oracle
- Issued on
PCAP: Programming Essentials in Python
- Issued by
- Cisco
- Issued on
Getting Started with Competitive Programming
- Issued by
- NPTEL
- Issued on