AI · Internal Tool
Intelligent Incident Triage Platform
LangChain · Llama 3.2 · Ollama · ChromaDB · Playwright
Local-LLM agent that ingests tickets, runs diagnostics, queries logs and DBs, and drafts RCA — cutting MTTR on production incidents.
I'm Harshal Wagh — a Software Developer at Amdocs with 3.5+ years driving quality engineering, system validation and automation for the Amdocs Billing Product supporting a major U.S. telecom provider, sustaining 99.95% billing accuracy in production.
I build and safeguard the systems that keep enterprise telecom billing running. My day-to-day blends Python, Shell and SQL automation with Linux systems work, self-healing frameworks and deep RCA — with a strong bias toward observability, repeatable releases and reducing manual toil.
At Amdocs I own validation and correction frameworks for the Amdocs Billing Product and lead AI-assisted tooling — including a local-LLM Incident Triage platform built on LangChain, Llama 3.2 and Playwright — that helps engineers resolve production incidents faster.
3.5+ years at Amdocs delivering automation, AI tooling and quality engineering for the Amdocs Billing Product supporting a major U.S. telecom provider.
A layered stack across automation, AI, cloud and backend engineering.
Self-healing frameworks, system validation and release readiness at telecom scale.
Local-LLM tooling that shrinks MTTR and automates triage across the stack.
AWS-certified with hands-on ELK, Grafana and production-grade Linux ops.
Billing workflows, storage & file systems, and CI/CD for enterprise delivery.
A mix of AI-assisted operations tooling, self-healing automation and billing-scale reliability engineering.
Local-LLM agent that ingests tickets, runs diagnostics, queries logs and DBs, and drafts RCA — cutting MTTR on production incidents.
Health-check + auto-remediation framework for Linux daemons and batch jobs, reducing manual debugging effort by up to 98%.
Automation-driven validation and correction frameworks that detect processing defects and data inconsistencies — sustaining 99.95% billing accuracy.
High-performance Shell automation migrating billing offers for millions of users, with pre/post safeguards guaranteeing 100% data integrity.
Regex + JavaScript pipelines on the ELK stack that surface anomalies in real time and route actionable alerts to on-call teams.
Internal tooling that aggregates defect lifecycle, RCA findings and environment health so release managers can green-light deployments faster.
Recognized for automation, AI/ML enablement and operational excellence.
Achieved 99.95% billing accuracy through automation and proactive error handling.
Implemented AI/ML-based configurations for early detection of system issues.
Integrated internal monitoring to improve anomaly detection and operational efficiency.
If you're building reliable platforms, automating away toil or bringing AI into engineering workflows — I'd love to talk.