Bryson Gracias
Open to work1 years experienceAI/ML Developer
AI/ML Developer specializing in Agentic AI, LLMOps, and RAG pipelines.
Standing Rank
Rank Not Available
Developer Badges
Skills & Technologies
71 skillsWork Experience
Agentic AI Engineer
Bayan DQ (Consultant)
- Architected a multi-agent system (Google ADK) translating natural-language queries into autonomous, multi-step Data Quality Assessment workflows for 100+ enterprise datasets.
- Streamlined user onboarding and execution workflows, accelerating platform operational efficiency by 70% and reducing average manual evaluation time by 30 hours per week.
- Deployed an MCP server for secure agent-to-server tool execution, automating ad-hoc data ops queries and reducing end-to-end processing latency by 40%.
Jr. Software Engineer
Germin8 Pvt Ltd
- Deployed a production LLM chat agent (Python, Gemini APIs) for automated CRM customer support, handling 15k+ monthly incoming conversations and reducing manual response formulation by 80%.
- Optimized response relevance to 94% accuracy via context engineering, RAG tuning, and structured NLP pre-processing for client-facing support conversations.
- Engineered real-time data ingestion pipelines (Apache Kafka, REST APIs, Webhooks), slashing event processing latency from 2.5 hours to sub-seconds while handling 100k+ events daily with 99% reliability .
Projects
GDELT & SEC based Graph-RAG Portfolio Agent
View Project- Developed a stateful Graph-RAG platform (LangGraph, Neo4j, OpenAI) for end-to-end stock analysis leveraging GDELT and SEC corpora.
- Designed an agentic ETL pipeline transitioning from raw data ingestion to Neo4j graph indexing and hybrid semantic retrieval.
- Instrumented distributed tracing (Langfuse) and structured JSON logging for LLM span-level observability and multi-step debugging.
- Built a FastAPI backend and Vis.js/D3 frontend for real-time knowledge graph visualization.
Meridian Customer Intelligence Platform
View Project- Engineered a hybrid predictive ML and generative RAG system for customer behavior analysis and automated support workflows with model monitoring and LLM tracing.
- Architected a stateful RAG pipeline (LangGraph, ChromaDB) featuring contextual grounding and explicit out-of-domain refusal mechanisms.
- Deployed via Azure Container Apps, FastAPI, and Nginx, delivering low-latency REST APIs for production client consumption.
FinOps-Copilot – Agentic AI Platform
View Project- Developed an agentic platform for NLP-driven browser automation featuring parallel batch execution and modular prompt workflows.
- Integrated real-time voice agents (LiveKit) for synchronous client interaction and automated query resolution.
RAG-DocQA – Production-Ready Multi-Document RAG Application
View Project- Engineered a multi-document RAG application utilizing semantic chunking and MMR reranking for high-precision retrieval across large corpora.
- Implemented automated testing and CI/CD pipelines (GitHub Actions, Jenkins) culminating in Dockerized deployment.
Cattle Detection System for Road Safety (YOLOv8)
View Project- Fine-tuned a YOLOv8 object detection model for real-time cattle detection, applying confidence thresholding to reduce false positives before triggering roadway alerts.
- Implemented zone monitoring and automated email alerts to notify authorities when cattle are detected within a designated area, supporting road safety and traffic management.
VaultGuard Operations Assistant
View Project- Architected a multi-agent operations assistant (CrewAI, FastMCP) automating research, compliance audits, and reporting.
- Implemented evidence-validation heuristics and prompt-injection filtering, mitigating hallucination and adversarial execution risks.
- Integrated dual-loop observability (Langfuse tracing, offline evaluation) for comprehensive audit trailing and regression prevention.
Instacart Product Stock & Reorder Prediction System
View Project- Deployed an end-to-end ML system forecasting consumer reorders (Macro F1: 0.82) to optimize FMCG inventory for 200K+ users.
- Engineered 20+ spatiotemporal and behavioral features modeling customer purchasing heuristics and routine patterns.
- Developed a multi-model pipeline (GPU-accelerated XGBoost, LightGBM) with automated model selection, yielding 5x training acceleration.
- Architected a scalable inference engine using FastAPI and Ray Serve for concurrent, low-latency batch and real-time predictions.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
Runner-up – MumbaiHacks 25
Runner-up – MumbaiHacks 25, World’s Largest Agentic AI Hackathon
Winner – National-level Generative AI Hackathon
Winner – National-level Generative AI Hackathon by Oraczen AI (Zenathon)
Winner – 24-hour Hackathon
Winner – 24-hour Hackathon organized with GDSC (GDG)
About Details
Professional Bio
AI/ML Developer with 1.5+ years proficient architecting ML, gen-AI agentic workflows across GCP, Azure, and AWS. Expertise in LLMOps, RAG pipelines, context engineering, and deploying scalable AI solutions to production environments.
Post Graduate Diploma in Artificial Intelligence, Machine Learning & Agentic AI in Artificial Intelligence, Machine Learning & Agentic AI
Indian Institute of Technology (IIT), Gandhinagar (2026 - 2026)
B.Tech in Computer Science and Engineering in Computer Science and Engineering
National Institute of Technology, Goa (2020 - 2024)