14K+
Daily deliveries supported
Enterprise Tookan dispatch workloads in production
MOHIT PANT
I design systems,
not just software.
Full-Stack Engineer working across backend architecture, cloud infrastructure, routing optimisation and AI-native developer tooling. Enterprise dispatch platforms supporting 14,000+ deliveries a day.
- Node.js
- TypeScript
- Python
- FastAPI
- AWS
- Docker
- Kubernetes
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Qdrant
- MCP
- Angular
- React
System blueprint
SELECT A NODE
CLIENT
Dispatcher consoles and operations dashboards
Angular v5 → v18 modernisation into microfrontends; operational efficiency up 45%, data accessibility up 60%.
Angular · React
Production impact
Systems worked at scale
Measured outcomes from production dispatch platforms and the routing engine. Every figure comes with the change that produced it.
2,000
Stops planned per routing request
Pickup-and-delivery routes under 12 hard constraints
143s → 30s
End-to-end VRP solve time
Typical 60–70 task workloads after pipeline rework
235s → 52s
Distance-matrix acquisition
Bounded worker pool replaced barrier-synchronised fetching
1,225 → 174
Provider calls per request
Hilbert-ordered spatial pruning cut redundant lookups
61.3%
API response-time reduction
Replication, read replicas and near-zero-downtime migrations
89%
Incident response-time reduction
Automated monitoring and alerting pipelines
257
Automated routing tests
GoogleTest unit, integration and endpoint suites, gating CI
Systems
What I've built and operated
Each one opens into a case study — the constraint it started from, the architecture, and what changed as a result.
Also built
EVOQ Antalia
- HTML5
- CSS3
- JavaScript
A broken multi-page real estate marketing site rebuilt as a single-page app with client-side routing across nine views — floor-plan tabs, a gallery lightbox, a custom SVG radial distance diagram, and lead capture with honeypot spam protection. Accessible focus states and reduced-motion support throughout.
evoq-antalia.mepant.com ↗Export tool
- JavaScript
- Node.js
Performance-based agent reporting built into the logistics backend, supporting incentive structures for delivery operations.
URL shortener
- MongoDB
- MySQL
- Node.js
Full-stack link shortener splitting storage by shape — MongoDB for analytics, MySQL for structured data — built with performance and scalability as the constraint.
Production systems
Where this ran
Enterprise dispatch work delivered on the Tookan ecosystem. Implementation detail stays with the clients; what follows is the shape of the work.
NEOM
SAUDI ARABIA
Smart city traveller & fleet dispatch
Vision 2030 gigaproject
Tookan validated and integrated as the dispatch engine for traveller and fleet logistics across NEOM zones and Neom Bay Airport.
- Real-time fleet tracking, automated task assignment and route-optimisation flows
- Custom APIs and webhooks for zero-manual-intervention workflows
- Production monitoring with automated alerting on SLA-threshold breaches
- Node.js
- MySQL
- Redis
- Angular
- AWS
Red Sea Global
SAUDI ARABIA
Tourism destination logistics
PIF-owned Vision 2030 gigaproject
Dispatch integration delivered and validated across 90+ islands and luxury resorts, targeting 1M+ annual visitors at full buildout.
- ACL/RBAC designed and validated for hierarchical teams
- Scheduling and dispatch automation reducing manual intervention
- Custom reporting and data-export tooling, with gap analysis for multi-destination decisions
- Node.js
- MySQL
- Angular
- AWS
DUBZ
DUBAI, UAE
Home baggage check-in, powered by dnata
10+ airlines at Dubai International
Dispatch platform architected and validated against Emirates flight schedules and SLA requirements for doorstep baggage check-in and delivery.
- Intelligent driver-assignment and route-clustering logic, stress-tested under peak load
- GPS tracking API integration with query and Redis-cache optimisation
- Zero missed SLAs through peak travel periods
- Node.js
- MySQL
- GPS APIs
- Angular
Additional engagements
Peddlers
AMSTERDAM
Rapid delivery
~14,000 deliveries/day
Owned and validated a Tookan-based delivery platform at roughly 14,000 deliveries a day.
Poona Cabs
INDIA
Employee transport scheduling
Automated employee transport scheduling customised and validated with route-distance optimisation.
McDonald's India
INDIA
Outlet maintenance scheduling
A Tookan fork built and validated for outlet maintenance scheduling with hierarchical access control.
Experience
The record
PROBLEM
Enterprise clients across Saudi Arabia, the UAE and Europe needed one configurable dispatch platform to behave like their own product, without the shared codebase fracturing.
ARCHITECTURE & IMPLEMENTATION
Platform capabilities decomposed into independently managed services, including a dedicated VRP engine. Client requirements translated into dispatch configurations, technical specs and validation criteria. Replication, migrations and monitoring added where the platform actually broke.
IMPACT
API response times down 61.3%, incident response down 89%, operational efficiency up 45%. Employee of the Quarter, twice.
PROBLEM
A link-shortening service that had to stay fast while recording enough data to be useful.
ARCHITECTURE & IMPLEMENTATION
Full-stack build splitting storage by shape — MongoDB for analytics, MySQL for structured data.
IMPACT
Shipped with performance and scalability as the design constraint.
IMPACT
GPA 8.4 / 10.
Method
How I approach systems
01
Understand the business constraint
The requirement is rarely the constraint. Find the one that actually binds.
02
Identify bottlenecks and failure points
Where does this break first, and what happens to everything downstream when it does.
03
Separate responsibilities
Split along the seams that let pieces be changed and deployed on their own.
04
Automate and observe
If it has to be watched by a person, it will eventually be missed by a person.
05
Measure the result
A change without a before and after is an opinion.
About
How I work
I build production systems that are scalable, observable and automation-friendly — and I spend most of my time on the parts that decide whether a system survives contact with real operations.
At Jungleworks I deliver and validate enterprise dispatch integrations on the Tookan ecosystem for NEOM, Red Sea Global and DUBZ across Saudi Arabia and the UAE, working directly with CEOs, CTOs, product stakeholders and engineering teams through full-cycle delivery. Outside that, I build independently: a pickup-and-delivery routing engine, and a multi-tenant SaaS platform.
- BACKEND ARCHITECTURE
- Service segmentation, configurable business logic, failure-point analysis, ACL/RBAC.
- DATA INTEGRITY
- MySQL replication and query optimisation, PostgreSQL, MongoDB, Redis, near-zero-downtime migrations.
- CLOUD & INFRASTRUCTURE
- AWS EC2, RDS and S3, Docker, Kubernetes, Traefik, HAProxy, Celery, CI/CD.
- AI & AGENT ENGINEERING
- Model Context Protocol tools, LLM fact extraction, shared agent memory on Qdrant.
- OPTIMISATION
- Adaptive large neighbourhood search, constraint modelling, spatial pruning.
- TESTING & VERIFICATION
- GoogleTest on the C++ engine, Mocha.js for automation workflow validation, Jest on platform code, Semgrep and ZAP in the pipeline.
- FRONTEND
- Angular v5–v18, microfrontends, React, REST integration.
FULL-STACK ENGINEER · BACKEND SYSTEMS, CLOUD & AI
Contact
Open to architecture, backend and AI engineering work
Fastest route is email. Happy to walk through any of the systems above in detail.