Amin
Fateminia
My first job was keeping servers running — Linux environments, network gear, on-premise VoIP for regional clients. Not the kind of work that comes with a clear playbook. You learn what reliability means when you're the one responsible for it, and when the call comes in at 2 AM.
From there I moved into VoIP engineering and PBX architecture — Asterisk, Issabel, custom dialplans, SIP trunk provisioning for organizations that needed real telephony infrastructure. That period built a working understanding of how voice networks actually behave: signaling, codec negotiation, the routing edge cases that only surface under real load. It's knowledge that still comes up in every cloud migration I work on.
Amazon Connect became the primary focus over time, once the platform was mature enough for serious production use. I've run the migrations, designed the contact flows, wired up the Salesforce integrations, and dealt with the problems that don't announce themselves until after go-live. The technical work is usually the easier half — the harder part is coordinating cutovers, setting realistic expectations, and making sure teams can operate the system after handoff.
I've worked across the US, Canada, Ireland, and Turkey. Different carriers, different compliance environments, different ways people think about communication infrastructure. It calibrates your assumptions in ways that are hard to replicate otherwise.
Outside work, I fly drones and play guitar — both of which require more patience than they appear to, and both have a way of clarifying what matters.
Some systems last.
Others reveal themselves at 3 AM.
Six areas.
Years of
specific depth.
I've worked in each of these areas long enough to know where they get complicated — the integrations that fail in non-obvious ways, the edge cases that only surface in production, the decisions that seem reasonable until you've lived with the consequences.
These aren't services off a checklist. They're the specific problems I've worked through repeatedly, across different clients, different operating environments, and different constraints.
Amazon Connect
Contact Center · AWSI've designed and deployed Amazon Connect environments across a range of scales — from straightforward routing setups to multi-site contact centers handling hundreds of agents across channels. The technical surface is broad: contact flow architecture, queue and routing logic, agent workspace configuration, real-time monitoring. The harder problems usually involve routing edge cases, failover behavior, and what happens when assumptions about call volume turn out to be wrong.
- Contact flow and IVR design across voice, chat, and task channels
- Queue-based and attribute-driven routing with fallback handling
- Agent workspace configuration and supervisor monitoring tooling
- CloudWatch dashboards, metric alarms, and operational alerting
- High-availability architecture with multi-region failover patterns
AI & Automation
Amazon Lex · Bedrock · LambdaThe promise of AI in contact centers gets overstated regularly. What actually works is well-scoped intent recognition with careful fallback handling — systems that know when to stop trying and transfer to a human, not just when they succeed. I design the automation layer with those exit paths as a first concern, not something to figure out after the demo.
- Amazon Lex voicebot and chatbot development with slot filling and session management
- Lambda-backed integrations for real-time account and order data lookup
- AWS Bedrock for generative response handling in appropriate contexts
- Escalation paths designed for clean, context-preserving handoff to agents
- Post-call analytics and contact classification pipelines
Salesforce Integration
CTI Adapter · Service CloudThe boundary between a contact center and Salesforce is where a lot of projects quietly stall — screen pops firing against the wrong record, call logs attaching inconsistently, click-to-dial that works in testing but breaks under concurrent load. I've worked through these problems enough times to know where to look before they become incidents.
- Salesforce CTI adapter configuration and troubleshooting
- Screen pop logic for inbound and outbound call scenarios
- Automated call log creation and case association workflows
- Click-to-dial integration from any Salesforce object type
- Lambda-to-Salesforce API integration and event-driven data sync
VoIP & PBX Engineering
Asterisk · Issabel · SIPBefore cloud contact centers were a viable option for most organizations, I was building Asterisk and Issabel environments for clients who needed real telephony infrastructure. That work developed a working understanding of how voice networks actually function — SIP signaling, PSTN interconnects, codec behavior under real conditions — that makes migrations more straightforward, because I understand what's being left behind.
- Asterisk and Issabel PBX architecture, deployment, and dialplan development
- SIP trunk configuration, failover routing, and carrier management
- AGI scripting and AMI integration for custom call handling logic
- FreePBX and FusionPBX administration and customization
- On-premise-to-cloud migration assessment and transition planning
WebRTC & Softphone
Browser-based Calling · Real-timeBrowser-based voice works well until it encounters real network conditions. STUN/TURN configuration, NAT traversal, codec mismatch, audio degradation under packet loss — these problems don't appear in controlled testing. I've debugged them under production pressure and know how to build softphone implementations that hold up outside ideal conditions.
- Amazon Connect Streams integration and softphone customization
- Custom WebRTC client development for browser-based calling
- STUN/TURN infrastructure configuration and network traversal
- Cross-browser audio quality troubleshooting and tuning
- Embedded calling experiences within web applications and CRMs
Cloud Infrastructure
Terraform · AWS · GitHub ActionsGood infrastructure doesn't require constant explanation. I work with Terraform and CloudFormation so environments are reproducible, deployments are predictable, and the next person who inherits the codebase can understand what's there without having to reverse-engineer it.
- Terraform module development for Amazon Connect and Lambda environments
- CloudFormation for multi-account AWS provisioning and stack management
- GitHub Actions pipelines for automated testing and deployment
- IAM policy design with least-privilege access patterns
- EventBridge and SNS-based event-driven automation workflows
A few things
I've helped
build.
Three projects I can talk about in detail — the constraints, the decisions, and what actually happened.
On-Premise PBX to Amazon Connect
Three offices, each running its own PBX setup with no visibility between them. Agents and supervisors had no real-time reporting, call data wasn't reaching Salesforce consistently, and the infrastructure team was spending significant time on maintenance that wasn't improving anything for the people actually taking calls.
Migrated to Amazon Connect in three phases — one office per cutover window — with Salesforce CTI integrated from day one so agents had consistent tooling across sites. Built CloudWatch and QuickSight reporting before the final cutover so supervisors had real-time visibility before the old system was decommissioned. The coordination across cutover windows was the harder part of the project.
Automated First-Touch for Routine Contacts
Agents were handling a large proportion of contacts that were essentially the same questions — order status, delivery estimates, return eligibility. Predictable, repetitive, and expensive to route to a human at scale.
Built an automation layer using Amazon Lex for intent handling, with Lambda functions querying the order management system directly. Most of the design work was about fallback — knowing precisely when the bot should stop trying and transfer to a live agent, and making that handoff carry enough context that the agent isn't starting from scratch. That's usually where these systems go wrong.
Unifying Two Separate Communication Environments
A Dublin office on Asterisk, a US team on a separate VoIP provider, and no shared customer context between them. Agents in each region were working from different systems, and there was no consistent way to see what was happening across both sites.
Built a hybrid setup that bridged the existing Asterisk environment into Amazon Connect over SIP — giving both teams a unified agent interface without forcing an immediate full migration. Added a contact data layer in S3 and Athena for cross-region reporting. The SIP interconnect required careful attention to codec alignment and routing behavior across two different network environments.
No handoff
gaps.
End-to-end.
From SIP signaling and carrier connections through to Lambda functions and CloudWatch dashboards — the full operational range of a contact center system.
- Amazon Connect
- Amazon Lex
- Amazon Polly
- Asterisk PBX
- Issabel / FreePBX
- FusionPBX
- SIP Trunking
- WebRTC
- RTP / SRTP
- PSTN / SS7
- Salesforce CTI
- VoIP
- AWS Lambda
- DynamoDB
- API Gateway
- EventBridge
- CloudWatch
- Amazon S3
- Athena
- SNS / SQS
- AWS Bedrock
- CloudFormation
- IAM
- Amazon Polly
- Terraform
- GitHub Actions
- Docker
- Linux (RHEL / Ubuntu)
- Python
- TypeScript / Node.js
- Bash
- REST APIs
- SQL
- Postman
Ten years.
Four countries.
The type of work changed over time. The underlying concerns — reliability, operational visibility, how systems behave under load — stayed the same.
Senior Amazon Connect Consultant
Independent · RemoteMost engagements at this stage are migrations and integrations — moving organizations from legacy telephony to Amazon Connect, usually with Salesforce in the mix and some level of automation on top. The work is technically broad but operationally specific: cutover coordination, contact flow design, agent tooling configuration, and making sure internal teams can own the system after handoff.
- Planned and executed multi-site migrations to Amazon Connect with coordinated cutover windows
- Designed Lex-based automation flows that reduced routine agent-handled contacts significantly
- Integrated Amazon Connect with Salesforce Service Cloud across multiple independent engagements
Cloud Communication Engineer
Consulting · US & CanadaThe shift from on-premise telephony to AWS-based contact centers was accelerating, and clients needed someone who understood both sides of that transition — not just the cloud target state, but the systems being replaced. Most of this period was migration planning, Lambda-backed routing logic, and helping teams understand the operational differences between managing hardware and working with a managed platform.
- Coordinated contact center migrations across business and technical stakeholders without service disruption
- Built Lambda-based routing and analytics pipelines integrated with existing business data systems
- Established Terraform patterns for reproducible multi-account AWS environment deployments
VoIP Engineer & PBX Architect
Consulting · Ireland & TurkeyAsterisk, Issabel, FreePBX — and the surrounding telephony ecosystem. This period developed a working understanding of how voice networks actually behave: SIP signaling, codec negotiation, PSTN interconnects, dialplan logic under real load. Working across Ireland and Turkey meant navigating different carriers and regulatory environments, which broadened my sense of how telecom infrastructure varies by geography.
- Designed and deployed Asterisk and Issabel PBX environments for organizations across multiple sectors
- Built WebRTC softphone integrations that replaced hardware phone dependencies for distributed teams
- Developed AGI scripts and AMI integrations for custom call routing and CRM connectivity
System & Network Administrator
Full-time · RegionalWhere it started. Linux servers, network infrastructure, on-premise VoIP — the kind of work where you learn what reliability means because you're the one who gets called when something breaks. That period developed habits around monitoring, understanding systems before touching them, and documenting what you've changed. Those habits have been useful through every role since.
- Managed Linux server environments and network infrastructure for regional organizations
- Deployed and maintained on-premise VoIP systems alongside network and security infrastructure
- Built monitoring and backup automation that reduced unplanned downtime and manual recovery work
Tuesday nights,
the rig is on.
A Strat plays through a tube amp on the same desk where, earlier that day, contact-flow JSON was getting committed. The pedalboard is small, opinionated, and over-engineered — which, if you know me, tracks.
→ click pedals to toggle
Sunday mornings,
the drone is up.
CAA-licensed (A2 CofC). Mostly coastal cliffs, open fields, the occasional cathedral. Flight logs live next to commit logs. Both teach you something about wind.
If you're working
on something
in this space —
If you're working through an Amazon Connect migration, a Salesforce integration that isn't behaving the way it should, or operational problems in a contact center environment — feel free to get in touch. I'm also happy to give a second opinion on architecture before decisions get locked in.