Ivan Shoolife
Kotlin
Python
Data Pipeline & Feature Engineering: 6/10
Active 18 days ago
Invite to interview
Message
Download CVCV
Overview
Technical skills
Roles
Overview
A backend-focused engineer at about a Senior level who authors production Python services and non-trivial domain logic. The strongest proven skill is building reliable document-consolidation backend logic and a file-backed registry - evidenced by contract_mvp/engine.py (consolidation algorithms) and contract_mvp/registry.py (transactional storage, revision management) with tests. Public code lacks evidence of multi-service distributed design, systematic migration history, or mature observability/ops integrations.
Technical skills
Kotlin• Middle
Python• Middle
Kotlin
Hilt
Kotlin Coroutines
Kotlin Flow
OkHttp
Retrofit
Python
FastAPI
Databases
Databases
SQLite
AI/ML
OpenCV
PaddleOCR
PaddlePaddle
NumPy
Pandas
Mobile
Android
Android SDK
Jetpack Compose
Material3
Room
SQLCipher
ViewModel
DevOps
Rest API
Middle AI/ML Engineer
Confidence: Medium LLM Engineer
A pragmatic engineer who builds production-ready document intelligence systems with reliable data pipelines, OCR/layout parsing and LLM-assisted extraction. The strongest proven skill is deterministic document consolidation and anomaly detection, evidenced by contract_mvp/engine.py and the registry lifecycle with tests in mvp/tests/test_registry.py. There is limited evidence of custom model training, experiment tracking or measured performance optimization in public code.
Model Architecture & Training
2/10
How well models are designed and trained
Integration-focused model engineering: code integrates LLM inference (HTTP/CLI) and uses pretrained CV/TSR models but contains no custom model architectures or training loops.
Data Pipeline & Feature Engineering
6/10
How data is prepared for models
Strong data pipeline and feature engineering for document/OCR processing - coordinated OCR, layout detection, table parsing, heuristics and deterministic consolidation logic.
Evidence
Shoolife/HackNado/mvp/contract_mvp/registry.py: registry, file storage, fingerprinting and transactional sqlite usage for document lifecycle
qSHEMq/RAG_PP/core/extractor.py: _build_grid_from_ocr_boxes/_build_grid_from_ocr_texts and many extraction heuristics
qSHEMq/RAG_PP/core/table_parser.py: parse_table_items - detailed column detection, numeric normalization and per-row corrections
Experimentation & Evaluation
2/10
How results are measured and tested
Basic but real testing and validation; there are unit tests and saved artifacts, but no structured experiment tracking, ablation studies or ML experiment management.
MLOps & Deployment
3/10
How models are shipped to production
Practical inference and serving engineering - Starlette/Gradio endpoints, local LLM/CLI fallbacks and retry logic; limited MLOps (no CI/CD, monitoring or model versioning shown).
Computational Efficiency
3/10
How efficiently computing resources are used
Some efficiency-minded engineering (lazy model init, image scaling, temp files), but no measured profiling, quantization experiments or GPU/batch-performance results included.
Research Depth & Innovation
3/10
Depth of research and new ideas
Good applied research-level heuristics and anomaly detection in document consolidation and OCR parsing, but not novel ML research or custom layers/papers reproduced.
Expertise
Document Intelligence & OCR• Middle
LLM• Middle
Industries
Professional Services• Middle
Financial Services• Middle
Technologies
OpenCV
FastAPI
Pandas
NumPy
PaddlePaddle
PaddleOCR
Recommendations
- Lead development of document-OCR+LLM extraction pipelines and production inference services (ingest, parsing, reconciliation, audit trails).
- Build LLM orchestration and extraction integrations (LLM-host fallback, retries, lenient parsing, offline CLI fallback) for enterprise document workflows.
- Implement robust MLOps around current pipelines - add experiment tracking, performance benchmarks, model/version management and monitoring.
- Extend anomaly-detection ruleset into a configurable rules engine and add legal/workflow connectors for human-in-the-loop review.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Backend Developer
Confidence: High API Engineer
A backend-focused engineer at about a Senior level who authors production Python services and non-trivial domain logic. The strongest proven skill is building reliable document-consolidation backend logic and a file-backed registry - evidenced by contract_mvp/engine.py (consolidation algorithms) and contract_mvp/registry.py (transactional storage, revision management) with tests. Public code lacks evidence of multi-service distributed design, systematic migration history, or mature observability/ops integrations.
API Design
5/10
How well APIs are designed
Practical, consistent REST API surface with defensive error handling and size/format checks, but no formal versioning scheme, no documented idempotency keys and only basic pagination/filters.
Data Layer & Database
6/10
Working with databases
Well-structured data layer using SQLite with explicit schema creation, indexes and transactional context management; good use of constraints and uniqueness but no multi-stage migration history is present.
Scalability & Performance
3/10
Handling load and speed
Some attention to concurrency and DB durability (WAL, locks) and non-blocking web handlers, but no queue-based decoupling, cache invalidation strategy, or explicit load/throughput design for high-scale workloads.
System Architecture
5/10
Overall system structure
Clear module boundaries (web API, registry, consolidation engine, service) and pragmatic file-backed job/revision model; architecture is suitable for a single-node service but not evidence of cross-service decomposition or distributed orchestration.
Security & Auth
6/10
Protecting data and access
Good input validation, path-safety checks and parameterized SQL use; careful handling of upload sizes and allowed suffixes reduces common injection/abuse vectors, though secret-management and auth lifecycle beyond basic session checks are not central here.
Reliability & Observability
6/10
Stability and monitoring
Defensive HTTP boundaries, explicit rollback on DB errors, and a decent test-suite increase reliability and observability; classic operational patterns (timeouts, retries with backoff, structured metrics, distributed tracing) are not present in the analyzed code.
Expertise
Python• Middle
Microservices & API Architecture• Middle
Industries
Professional Services• Middle
Technologies
Databases
Python• Middle
Rest API
SQLite
Core• mentioned only
Tauri• mentioned only
Recommendations
- Design and implement backend services that ingest and consolidate domain documents - e.g., legal document processing pipelines and registries (use the consolidation engine and registry patterns shown).
- Build REST APIs that require careful file handling, validation and revision history - own endpoints, upload validation, and file-backed job models similar to web.py and ContractRegistry.
- Harden operational aspects: add migration tooling, structured logging/correlation ids, metrics and retry/backoff policies for long-running consolidation jobs.
- Extend the architecture to a worker-queue or background-job system (Celery/RQ/Task queue) for larger scale ingestion and safe retries, reusing the existing service.run and revision registration patterns.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Mobile Developer
Confidence: Medium Android Engineer
An experienced Android engineer focused on Android TV platform work, at a senior level for native platform integration and concurrency. The strongest proven skill is building robust platform services and networking logic - evidenced by VpnConnectionManager.kt (generation counters, mutexes, health checks) and ToBeVpnService.kt (VpnService lifecycle and foreground-service handling). Public code lacks measured performance benchmarks, CI/release automation and a full offline-sync engine with conflict-resolution.
Platform Native Mastery
7/10
Knowing the mobile platform
Strong Android-native systems work: service lifecycle, foreground service start handling, VpnService integration, structured concurrency with cancellation, mutex-protected state and generation counters to avoid stale updates.
Evidence
app/src/main/java/com/tobevpn/tv/vpn/VpnConnectionManager.kt: handleServiceDestroyed/updateState/startVpnInternal with generation counters and mutex
app/src/main/java/com/tobevpn/tv/vpn/ToBeVpnService.kt: VpnService subclass, notification / service interactions
app/src/main/java/com/tobevpn/tv/vpn/XRayCore.kt: native core integration (libv2ray/go.Seq) and stat querying
Mobile UI/UX & Responsiveness
6/10
Smooth mobile experience
Well-executed Compose TV UI with explicit focus/navigation handling, scaling for large screens, D-pad focus properties and lifecycle-aware state collection - targeted at Android TV UX rather than generic templates.
Evidence
app/src/main/java/com/tobevpn/tv/presentation/referrals/ReferralsScreen.kt: detailed focus requesters, D-pad navigation and TV-adapted layout/scaling
app/src/main/java/com/tobevpn/tv/presentation/subscription/SubscriptionScreen.kt: tariff tabs, focus restoration, QR overlay and Compose performance-conscious code
app/src/main/java/com/tobevpn/tv/presentation/TvScale.kt: screen scaling utilities for TV form-factors
Performance & Battery
6/10
Speed and battery use
Performance and battery-awareness are present: background/foreground service handling, heartbeat and probe strategies, health checks and NonCancellable accounting to avoid data loss - but no published measurements or profiles.
Evidence
app/src/main/java/com/tobevpn/tv/vpn/VpnConnectionManager.kt: heartbeat (HEARTBEAT_TICKS), tunnel health checks and probeTunnelWithRetries
app/src/main/java/com/tobevpn/tv/update/UpdateDownloader.kt: concurrent download tasks, OkHttp usage and cancellation-aware flows
app/src/main/java/com/tobevpn/tv/vpn/VpnConnectionManager.kt: careful use of Dispatchers.IO, SupervisorJob, and NonCancellable for flushPendingUsage/drainTrafficCounters
Offline & Data Sync
4/10
Working offline and syncing
Some offline-first and local persistence work exists (usage accounting, session logs, traffic logging), but there is no full-featured sync engine with conflict resolution or explicit queued idempotent retry semantics visible.
Evidence
app/src/main/java/com/tobevpn/tv/data/local/ (Room daos referenced) and app/src/main/java/com/tobevpn/tv/di/DatabaseModule.kt: Room + SQLCipher integration
app/src/main/java/com/tobevpn/tv/vpn/VpnConnectionManager.kt: saveSessionLog/flushPendingUsage/drainTrafficCounters - local traffic accounting
Device Integration
6/10
Using device features
Solid device and platform integration: VpnService, Notifications, PendingIntents, FileProvider/installer flow, and handling of Android 12+ background service start restrictions and network capability checks.
Evidence
app/src/main/java/com/tobevpn/tv/vpn/ToBeVpnService.kt: Notification, PendingIntent and VpnService usage
app/src/main/java/com/tobevpn/tv/update/UpdateInstaller.kt: FileProvider, unknown-sources checks and install Intent handling
app/src/main/java/com/tobevpn/tv/vpn/VpnConnectionManager.kt: hasUnderlyingInternet(), isOwnVpnNetworkActive() and connectivity capability checks
Release & App Lifecycle
3/10
Building and publishing apps
Basic release and lifecycle hygiene (in-app update checks, GitHub releases API usage) is present but there is no evidence of CI/CD pipelines, fastlane, signing automation or staged rollout / crash triage configs in public code.
Evidence
app/src/main/java/com/tobevpn/tv/di/NetworkModule.kt: GithubReleasesApi and update check wiring
app/src/main/java/com/tobevpn/tv/update/UpdateBannerCheck.kt and UpdateDownloader.kt: built-in updater components
Verified artifacts
Expertise
Android• Middle
Mobile Performance & Optimization• Middle
Mobile UI/UX• Middle
Mobile Security & App Integrity• Middle
Industries
Consumer Electronics• Middle
Internet Services• Middle
Technologies
Android
Kotlin• Middle
Jetpack Compose
Android SDK
Retrofit
Hilt
Room
Kotlin Coroutines
Kotlin Flow
SQLCipher
OkHttp
ViewModel
Material3
Recommendations
- Lead development of Android TV apps that require complex system integration - VPNs, media devices or IoT companion apps with foreground services and native libs.
- Implement concurrency-critical networking components and background-task orchestration (health checks, watchdogs, usage accounting, and resilient service restart).
- Design and harden in-app update and installer flows, plus native integration for performance-sensitive subsystems (JNI/gomobile) where careful lifecycle handling is required.
- Build polished TV-first Compose UIs with D-pad / focus navigation and accessibility for leanback devices.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: