Available for new opportunities

Shahul Hameed
Akmal

Building AI systems that scale. 4+ years shipping production LLM pipelines, vector search engines, and full-stack platforms that process hundreds of thousands of records.

View Projects Get in Touch
0
Records Ranked
0
Search Precision Lift
0
Answer Accuracy
0
Msg Delivery
0
Experience
scroll

About Me

AI Engineer.
Systems Thinker.

I'm a full-stack software engineer with a deep focus on AI/ML systems. I've spent 4+ years building production-grade LLM platforms, vector search engines, and data pipelines that operate at real scale.

At HiAcuity, I grew from intern to software engineer — owning end-to-end slices of the stack, from LangGraph AI agents and AWS infrastructure to React frontends and Stripe billing systems.

I care about engineering quality: observable systems, clean abstractions, and code that the next engineer can actually understand and extend.

📍 Jamaica, New York 🎓 B.Sc. CS · First Class Honours 🏢 HiAcuity · Sept 2022 – Present
🤖
AI & LLMs
LangGraph agents, RAG pipelines, vector search across 100K+ records at sub-second latency
⚡
Backend Scale
FastAPI, PostgreSQL, MongoDB — high-throughput pipelines, 30%+ query improvements
☁️
Cloud & DevOps
AWS-native (SQS, OpenSearch, ECS), Kubernetes, Terraform — cut infra costs 15–20%
🎨
Frontend
React, TypeScript, Next.js — 25% render reduction via architectural patterns

Career

Experience

Three years at the same company — intern to engineer, owning progressively more of the stack.

Software Engineer
HiAcuity · Remote (Australia)
Sept 2024 – Present
  • AI-Powered Data Assistant (LangGraph) — End-to-end analytics assistant enabling natural language data queries via dynamic query generation. AWS SQS event pipelines with multiple workers + DLQs; 99.5% delivery reliability, cutting review time by 40%
  • Semantic Search Engine (AWS OpenSearch) — Vector-based candidate matching lifting precision by 35% over keyword search; scaled to 100K+ profiles with sub-second query latency
  • Analytics Platform — Recruiter-facing ingestion pipeline on MongoDB + PostgreSQL; redesigned indexes improving query time by 30%+, enabling 2× shortlisting pace
  • React Architecture — useReducer + useContext + custom hooks cut component re-renders by 25%; built reusable data visualization dashboard components
  • Stripe Billing System — 3 pricing tiers, webhook-driven access control, reduced manual billing exceptions by 90%
  • Mentorship — Code reviews + pairing cut PR cycle time by 20%
Junior Software Engineer
HiAcuity · Remote (Australia)
Mar 2023 – Sept 2024
  • Microservice Standardization — Led cross-service refactor across 6 microservices, eliminating 1,500 lines of duplicate code and cutting engineer onboarding time by 50%
  • Cloud Cost Optimization — Used Prometheus + Grafana to right-size EC2 instances and redesign auto-scaling policies; cut monthly infra spend by 15–20% without degrading uptime
Software Engineering Intern
HiAcuity · Remote (Australia)
Sept 2022 – Mar 2023
  • Secure API Development — FastAPI endpoints with role- and group-based access control for two user personas; zero access violations in security audit
  • Test Automation — 40+ Selenium test cases covering critical UI flows; reduced manual regression effort by 60%; ran JMeter + Datadog load tests identifying 3 critical bottlenecks pre-production

Work

Projects

Proof-of-concept to production — systems that actually work at scale.

01
AI Data Assistant
Natural Language Analytics Engine

Multi-stage LangGraph orchestration pipeline translating natural language questions into structured database queries against PostgreSQL and MongoDB. Engineered a ReAct-style tool-calling loop with retrieval-augmented generation for self-serve analytics without SQL knowledge.

87% top-3 accuracy10,000+ records
LangGraphLangChainLlamaIndexPostgreSQLMongoDBOpenAI APIPython
02
AI Road Trip Planner
Conversational Agentic Route Generator

Voice-enabled conversational AI agent translating natural language trip descriptions into fully resolved multi-stop driving routes. GPT-4o autonomously selects and chains 5 Google Maps tools via ReAct loop. A 7-node LangGraph StateGraph manages conversation flow; OpenAI Whisper handles full voice I/O.

7-node StateGraphFull voice I/O5 tool chains
LangGraphGPT-4oWhisperGoogle Maps APIFastAPIReactTypeScript
03
AI Resume Search
Candidate Matching Engine

Multi-stage retrieval pipeline using LangGraph to orchestrate Pinecone vector search with structured MongoDB filters across experience, skills, and location. Achieves high match accuracy across a large dataset of resumes with multi-dimensional hybrid filtering.

87% top-3 match10,000+ resumes
LangGraphPineconeLlamaIndexMongoDBPythonLLMs
04
HouseDiffusion
Generative Floor Plan Generator

Trained a Diffusion + Transformer model for room-level layout synthesis, reducing floor plan generation from hours of manual drafting to under 30 seconds. Users interact via a React drag-and-drop canvas; FastAPI handles model inference serving.

<30s generationvs. hours manually
PyTorchDiffusion ModelsTransformersReact FlowTypeScriptFastAPIPython

Stack

Skills

Breadth across the stack, depth where it counts.

AI / ML
LangGraphLangChainLlamaIndex RAGVector SearchLLMs AI AgentsOpenAI APIAnthropic API PineconeHugging FacePrompt Eng.
Backend
PythonFastAPIPostgreSQL MongoDBGraphQLREST APIs SQLAlchemyPydanticNoSQL
Frontend
ReactTypeScriptNext.js Vue.jsTailwind CSSJavaScript StorybookWebpackESLint
AWS
EC2S3SQS SNSSESLambda OpenSearchECSIAM AmplifySageMakerCloudWatch
DevOps
KubernetesDockerHelm TerraformGrafanaPrometheus GitHub CI/CDCircleCINginx

Contact

Let's Build Something

Open to full-stack and AI engineering roles. I love hard problems and teams that actually ship.

s.hameedakmal@gmail.com
LinkedIn GitHub 203-997-4545