Open to new roles · available from October 2026

Haifa area, Israel (UTC+2/3) · full EU overlap

Beniamin Levin Lead Software Engineer — still hands-on after 19+ years.

I design, build, and lead the backend platforms behind enterprise telecom OSS/BSS — order entry, CPQ, product catalog — for operators including Telenet, Telefónica/O2, and Rakuten, after earlier years in automotive and mobile. Nineteen years in, I am still the person tuning the SQL and reading the thread dump: Java/Spring on Kubernetes and OpenShift by day, MCP- and RAG-grounded agents that cut our defect triage from two hours to thirty minutes, and a Go system I wrote and operate myself in production.

Java · Spring Boot 3 · Kubernetes / OpenShift · Go · MCP- and RAG-grounded agents

Order-entry latency
~4s → ~0.5s
Order Capture defects
−75% / 2 yrs
Defect triage
2h → 30m
Telecom customer programs
30+
Beniamin Levin — Lead Software Engineer

Selected Engineering Impact

Six outcomes I owned end to end, with the scope they were measured in. Each one has a measurement method and a set of trade-offs I am happy to walk through.

  • ~4s → ~0.5s

    Order-entry latency

    30–88% faster across critical enterprise order-entry workflows — redundant service calls eliminated, targeted caching introduced, SQL and application hot paths tuned.

  • −75%

    Incoming defect volume

    Order Capture component, over two years. Recurring root causes identified and eliminated, code review and CI validation strengthened, quality fixed before changes reached customer environments.

  • 2h → 30m

    Average defect-triage time

    A Claude Sonnet ticket-assessment agent for Order Capture. I authored its domain skill; every assessment is grounded through MCP tools (Jira, GitLab, Jenkins) and RAG retrieval over product documentation and prior tickets, so root-cause and fix suggestions cite their evidence.

  • −40%

    Unit-test creation time

    Repository-aware AI workflows rolled out across Java/Spring and Angular teams for implementation, refactoring, code analysis, and test generation.

  • 11 / 30+

    Team and delivery scope

    Led a distributed team of up to 11 engineers across Europe and India while staying hands-on, as technical owner across 30+ telecom customer programs and concurrent releases.

  • 70+ → 5

    Pune engineering pod

    A five-engineer pod built in three months: the team selected from 70+ interviews, then two multi-month on-site architecture and knowledge-transfer engagements to accelerate ramp-up.

Professional Experience

Nineteen-plus years building, scaling, and supporting enterprise telecom OSS/BSS, automotive, and mobile platforms — leading distributed teams while staying in the code.

Netcracker Technology

July 2015 – September 2026

Lead Software Engineer

Enterprise Telecom OSS/BSS Platform Architecture & Delivery

Java · Spring Boot · Kubernetes / OpenShift · Microservices · MCP / RAG Agents
  • Reduced response times by 30–88% across critical enterprise order-entry workflows, accelerating key flows from ~4s to ~0.5s, by eliminating redundant service calls, introducing targeted caching, and tuning SQL and application hot paths.
  • Led technical design, feature development, and scaling across enterprise telecom OSS/BSS platforms, spanning order entry, CPQ, product catalog, validation, and Kubernetes-based delivery streams.
  • Decreased incoming defect volume for the Order Capture component by ~75% over two years by identifying and eliminating recurring root causes, strengthening code reviews and CI validation, and improving engineering quality before changes reached customer environments.
  • Resolved 20 accessibility-related defects for Charter Communications.
  • Cut average defect-triage time from ~2 hours to ~30 minutes by operationalising a Claude Sonnet-based ticket-assessment agent for Order Capture — authored its domain skill, grounding each assessment through MCP tools (Jira, GitLab, Jenkins) and RAG retrieval over product documentation and prior tickets for evidence-cited root-cause and fix suggestions.
  • Scaled AI-assisted development across Java/Spring and Angular teams, introducing repository-aware workflows for implementation, refactoring, code analysis, and test generation — cutting unit-test creation time by ~40%.
  • Served as hands-on technical owner across multi-release OSS/BSS delivery for 30+ telecom customer programs (including Telenet, Telefónica/O2, GCI Alaska, GTD Chile, and Rakuten), resolving architecture, integration, performance, security, backport, and production-release issues across enterprise order-entry platforms.
  • Led a distributed engineering team of up to 11 developers across Europe and India while remaining hands-on, delivering concurrent customer releases and Level-3 production support while owning technical quality and critical incident resolution.
  • Built a 5-engineer Pune engineering pod in three months, selecting the team from 70+ interviews and accelerating technical ramp-up through two multi-month on-site architecture and knowledge-transfer engagements.

Symphony Teleca Corporation (now Harman)

February 2011 – July 2015

Principal Software Engineer

Automotive Infotainment, Mobile Platforms & Core Messaging

Java · Android · Connected Vehicle / Infotainment
  • Samsung Messaging (Suwon, South Korea): Selected for an on-site engagement at Samsung HQ to optimise and customise core messaging infrastructure for flagship Android devices.
  • SiriusXM Connected Vehicle Platform: Engineered the Java server-side architecture and Android client for a vehicle infotainment platform demonstrated live at CES 2012.
  • Android UI & Widgets (Nokia, Savelli): Developed low-level Android UI adaptations replicating native Nokia experiences, and built weather/alert widgets focused on responsiveness, asynchronous loading, and low power consumption.

Onlyoffice (TeamLab)

November 2009 – January 2011

Software Engineer

Enterprise Collaboration & Document Management Platform

C# · ASP.NET · SQL · LDAP
  • Designed and implemented operational subsystems in C# and ASP.NET for TeamLab, now Onlyoffice, an enterprise collaboration and document management platform.
  • Delivered modules for bookmark systems, corporate planners/organisers, project templates, real-time analytics aggregators, and automated bulk LDAP/employee data ingestion pipelines.

MERA NN

November 2006 – November 2009

Software Engineer

Telecom Communication Infrastructure Management

Java · JUnit / Cobertura · Nortel OA&M · High Availability
  • Built high-availability web applications for operating, remotely administering, and monitoring Nortel ICP communication infrastructure.
  • Authored architecture specifications and deployment plans; maintained JUnit/Cobertura test suites to raise coverage and reliability.

Selected Projects & Freelance

Systems I own outright — designed, written, deployed, and operated by me. ikite.fyi is public and inspectable if you want to see current Go rather than take my word for it.

Sure Communications

2019 – Present

Part-Time Freelance Software Engineer

Telecom & Web Integration Solutions for Business Clients

PHP · MySQL · AI / Grok API · OpenAI · Chart.js · Shopify

Part-time, evenings and weekends, outside full-time employment and in a non-competing domain.

  • Delivered end-to-end PHP/MySQL/JavaScript web applications, customer portals, CRM integrations, and back-office tooling for telecom and business clients, including Shopify e-commerce integrations, Memento platform delivery, and the flagship dealer/consumer sales platform for Morgan Motor Company (morgan-sales.com).
  • Led technical architecture, backend automation, and performance tuning across enterprise platforms (crm.sur.co.uk, sur.co.uk, altis.co.uk, melodycare.co.uk); built file-rendering services, cron task runners, email-to-comment synchronisation, and database tuning.
  • Engineered an OpenAI ChatGPT-powered email-reply generator (context-aware drafting with greeting personalisation, thread trimming, and chase-email detection) and an AI ticket comment generator drafting support responses from full conversation history.
  • Architected and built "SurAI": an embedded natural-language business-intelligence copilot in the CRM that converts plain-English questions into schema-aware SQL via the Grok (xAI) API. Injects metadata to prevent hallucinated fields, enforces read-only connections & SELECT-only guardrails, and renders results into interactive tables and Chart.js visualisations (bar/pie/line).

ikite.fyi

2023 – Present

Case study · sole engineer & operator

Live at ikite.fyi

Israeli Mediterranean wind & meteorological monitoring platform

Go · MySQL · OCI · systemd · Docker · ApexCharts · Telegram API

Kitesurfers on the Israeli coast decide whether to drive an hour based on wind that changes in twenty minutes. Public sources are scattered, slow, and mostly Hebrew PDFs. I built the thing I wanted to exist, then kept it running for three years.

Architecture

4 external sources → scheduled Go collectors (systemd timers) → MySQL time-series → Go HTTP API → ApexCharts UI + Telegram alerting

Collectors poll Windguru, the IMS Envista API, and local Israeli station feeds on independent schedules, so one flaky upstream cannot stall the others. A second pipeline ingests daily forecasts: openWRF PDF parsing, OpenSkiron 4 km GRIB, Open-Meteo models, and a Hebrew AI forecast report that is machine-translated to English on ingest. Everything runs on a single Oracle Cloud instance under systemd, with the binary and its assets deployed as one artifact.

Why I rewrote it from PHP to Go

The 2023 PHP version worked, but every collector was a separate cron-invoked process paying full interpreter startup, and concurrent polling meant either blocking or process sprawl. Go gave me goroutine-per-collector concurrency in one supervised process, a single static binary to deploy, and a memory profile that fits comfortably inside a free-tier VM. The 2026 rewrite is the version running today.

  • ~15 MB resident memory for the whole service, versus a PHP-FPM pool
  • Sub-millisecond API responses on cached station reads
  • One artifact to deploy — static binary plus embedded assets, no runtime to keep patched
  • Zero unplanned outages attributable to collector concurrency since the rewrite
What running it for three years taught me

Upstream feeds fail in undignified ways — a station returns yesterday's reading with today's timestamp, a PDF layout shifts, a GRIB file arrives truncated. Most of the engineering is not the happy path; it is staleness detection, per-source backoff, and alerting that distinguishes "no wind" from "no data". The Telegram threshold alerts had to learn the difference the hard way.

Business FM — Custom CRM

2012 – 2015

Full-Stack Architect & Developer

Custom Media & Radio Sales CRM

PHP · MySQL · AWS (EC2/RDS) · Microsoft Azure
  • Designed, built, deployed, and maintained a bespoke PHP/MySQL CRM for client, contact, and media sales-workflow management as the sole developer.
  • Initially deployed on Microsoft Azure before migrating production infrastructure to AWS EC2/RDS for improved cost optimization and operational control.

Core Competencies & Technical Skills

The same list as my CV, grouped the same way: the columns I work in every day come first, then the wider stack I have shipped with.

Technical Leadership & Delivery

Engineering Leadership System Design SOLID Principles Code Review Pull Request Review Distributed Systems Team Mentoring Agile / Scrum Sprint Planning Story Mapping L3 Production Support

Backend

Java Spring Boot 3 Spring Framework 6 Spring Cloud JPA Quarkus Tomcat Node.js REST SOAP GraphQL Microservices Go C# ASP.NET

Messaging & Data

Elasticsearch PostgreSQL RabbitMQ Redis Kafka Cassandra MySQL SQL NoSQL

Cloud & DevOps

Docker Kubernetes OpenShift GCP AWS (S3, EC2) Azure Oracle Cloud Linux Git CI/CD Jenkins Prometheus Grafana Monitoring, Logging & Telemetry

Frontend & Mobile

Angular JavaScript TypeScript jQuery HTML5 CSS3 Bootstrap PHP Android

AI-Assisted Engineering

Claude Code Cursor Antigravity Claude Sonnet Gemini OpenAI GPT Grok (xAI) API MCP Agent Tooling RAG-Grounded Agents Natural-Language-to-SQL LLM Prompt & Agent-Skill Engineering AI-Assisted Defect Triage Repository-Aware Code Analysis

Education & Professional Development

Academic foundation, the corporate training worth listing at this level, and working languages.

N. I. Lobachevsky State University of Nizhny Novgorod

2003 – 2008

Graduated with Honours

Master's Degree, Faculty of Computational Mathematics & Cybernetics

Nortel Networks

March 2009 – June 2009

New York, USA

Advanced Corporate Training: Common OA&M Architecture

Operations, administration, and maintenance architecture for carrier-grade telecom infrastructure.

Selected Professional Development

Engineering leadership, delegation, and multi-site team management
Microservices with Spring Boot and Spring Cloud
Docker, Kubernetes, and OpenShift
Kafka, RabbitMQ, Redis, and Cassandra
Angular architecture and state management

Languages

  • English

    Full professional working proficiency

    19+ years of daily use in international engineering teams, technical documentation, code review, and customer-facing delivery — including on-site work in India and South Korea and corporate training in New York, USA.

  • Hebrew

    Upper-intermediate — professional working proficiency

  • Russian

    Native

Open to new roles · from October 2026

Let’s Connect

I am available from October 2026, when my Netcracker role wraps up. I’m most effective in roles that combine hands-on backend engineering with technical leadership — the kind where the person setting the architecture is also the one who has to live in the codebase. Happy to talk about distributed system design, performance work, or getting MCP- and RAG-grounded agents usefully into an engineering team’s workflow.

Send Email LinkedIn
✓ Email copied to clipboard