Nexus
Turns plain-English business questions into validated T-SQL through a 17-node DAG workflow, then renders the answers as live dashboards and scheduled reports.
AI/ML Software Engineer. I build it, ship it, and put it online. Retrieval, agents and evals that hold up outside the demo.
PROJECTS
Turns plain-English business questions into validated T-SQL through a 17-node DAG workflow, then renders the answers as live dashboards and scheduled reports.
Dense vector search and BM25 fused by Reciprocal Rank Fusion, cross-encoder reranked, and gated in CI by 60+ automated evals so a regression never ships.
Every voice turn is intent-classified and routed to RAG, a tool call or a direct answer, holding sub-second audio latency while tracking p50/p95/p99 and cost per stage.
A small model taught to emit schema-valid JSON through QLoRA adapters and DPO preference alignment, at a fraction of full fine-tuning cost.
Fifteen cloud builds end to end: ALB/NLB autoscaling, multi-tier VPC deployments, serverless pipelines and automated cost optimisation, all provisioned as infrastructure-as-code.
An 8-node LangGraph that prospects a niche, crawls and researches each company, extracts C-suite contacts, then halts for human approval. It's checkpointed, so a run survives a restart.
JUNIOR AI ENGINEER, MAXNET TECHNOLOGIES MAR 2025 TO SEP 2025
TECHNICAL CONTENT CURATOR, MYEQUATION MAR 2025 TO JUN 2025
OPEN SOURCE
contributing to open source, here's my github streak and pull request activity.
CONTRIBUTION GRAPH
Loading live GitHub data…