AI/ML Engineer based in Seattle, WA. I build production AI systems — from Vision-Language Model pipelines and RAG architectures to multi-agent workflows and distributed inference infrastructure. Currently open to full-time AI/ML Engineer roles.
📍 Seattle, WA | 🌐 srikeerthi.dev | 💼 linkedin.com/in/srikeerthis
AI/ML & GenAI: RAG Pipelines, LLM Orchestration, LangChain, LangGraph, DINOv2, GLM, InternVL, YOLO, PyTorch, TensorFlow, Hugging Face
Vector Search & Data: Milvus, Faiss, LakeFS, MongoDB, Semantic Search, ETL Pipelines, OpenCV
Infrastructure: AWS (Lambda, EC2, SQS, S3, CloudWatch), Docker, Kubernetes, Kafka, Spark Streaming, GCP, CI/CD (GitHub Actions), Grafana, PostgreSQL
Languages & Frameworks: Python, FastAPI, Node.js, JavaScript, SQL, Django
Python · Gemini API · OpenCV · FastAPI · LangGraph · Three.js · SQLite
Built a three-agent AI pipeline for safe human-robot workspace collaboration. Gemini Vision + OpenCV detects objects at pixel precision; a hard geometric safety layer mathematically enforces human zones independent of model behavior; a self-improving planner injects rated past sessions as few-shot context to improve plan quality without retraining.
🏆 TechEx Intelligent Enterprise Hackathon — Track 3: Robotics & Simulation
Kafka · Spark Structured Streaming · Kubernetes · Grafana · FastAPI · PostgreSQL
Distributed observability pipeline for real-time windowed aggregations across 10+ simulated servers. Full alerting via Grafana dashboards and Discord integration, containerized with Docker and K8s StatefulSets for persistent storage.
Python · LinUCB · FastAPI
Distributed RTB microservice using LinUCB Contextual Bandits, achieving 73.90% conversion rate in high-frequency marketplace simulations. Inference path optimized to <20ms with asynchronous weight updates via FastAPI Background Tasks.
AI Engineer Intern — Olympic Collectibles AI Solutions(12/2025 – 03/2026)
Deployed VLM evaluation framework comparing GLM-4.6V-Flash vs InternVL 2.5 in production; built distributed inference system over 450GB / 150K+ image dataset at 99.15% retrieval precision.
Software Developer Intern — ecohome.one LLC(01/2025 – 09/2025)
Serverless event-driven backend on AWS Lambda + SQS with fault-tolerant DLQ pipeline and zero data loss architecture.
Machine Learning Engineer — Scientist Technologies(05/2021 – 07/2022)
Full ML lifecycle management for ResNet50 object detection; diagnosed 9% accuracy drift and deployed retrained model at 95% accuracy.
- 🥇 Smart India Hackathon 2019 — Winner
Built SmartLoad, a Unity3D application that displays optimized 3D cargo loading patterns for trucks using advanced bin-packing algorithms — maximizing space utilization and reducing planning time and operational costs. Recognized at India's largest national hackathon for its practical impact on logistics optimization.
- Participated in Hacktoberfest, Open Source Day, and Common Voice Sprint
- Contributed to Open Source Lab (VVCE) and Mobi at UTA
- Organized workshops promoting open-source development and social coding
Open to collaborating on AI/ML projects — feel free to reach out!