Currently · Freelance software engineer, building applied AI projects
Omar Salama
React Native & Python. RAG pipelines, on-device LLMs, five years of shipped products.
Open to remote (CET/EU) & relocation · Giza, Egypt
Two tracks, one engineer
Software engineer, five years in, working where AI systems and mobile overlap.
On the AI side: LLM systems end to end, from cited answer pipelines built on agentic tool-calling to retrieval with hybrid search and reranking, fine-tuning with SFT and DPO, and the evaluation tooling that keeps all of it honest. On the mobile side, native Android led me to React Native and years of shipped cross-platform features, most recently a fine-tuned LLM running fully on-device in a React Native app.
3.5 years at Keepoala, a year-plus LLM-training contract at Turing, then SDE II at Niyyah (2025–2026) across the AI backend and React Native app. Since June I have been freelancing: client app work alongside applied-AI projects such as the Campaign Review Copilot and Jetty. On the side, OpenSpace: an on-device therapy companion running a Gemma-4 E2B model I fine-tuned and published as TheraSpace on HuggingFace.
“The work I like best runs the whole length of a feature: the model, the retrieval that feeds it, and the app it ships in.”
Where I've worked
Jun 2026 - Present
Freelance
Software Engineer
Client app work alongside independent applied-AI projects.
- Client work: a white-label hotel guest app, web first and then iOS and Android, in English and Arabic with full RTL. Guests raise a service request and track its status; the app writes that request into the hotel's existing operations system through the client's API, plus the guest endpoints I agreed with their backend team.
- Built the Campaign Review Copilot, an AI triage prototype for a crowdfunding Trust & Safety team: every flag cites its FATF, OFAC, or fraud-taxonomy source, and a synthetic eval targets cultural false positives.
- Built Jetty, a voice travel assistant: it will not speak a visa rule, hotel price, or flight fare it did not just look up, because deterministic code fills each value from that turn's tool result or holds the line back.
Sep 2025 - Jun 2026
Niyyah
SDE II
Worked across the AI backend and the React Native app of a consumer Muslim app with 100K+ downloads.
- Built the AI backend's answer pipeline: agentic tool-calling with budgets and guardrails backs every cited, streamed answer with ranked sources, with multi-provider LLM fallback for quality, cost, and reliability.
- Built the evaluation tooling from scratch: an LLM-as-judge harness, retrieval-quality metrics, and regression checks with alerting that gate what ships, plus a plan agreed with the CEO and a domain expert to calibrate the judge against an expert-labelled golden set.
- Cut input tokens per answer 10–20%, worked on conversation summarization to keep long chats within budget, and built a server-side streaming proxy with stream resume, which fixed answers dropping mid-stream when the app backgrounded or the connection dropped.
- Architected the multi-agent flow behind a personalized 30-day guided journey, which the team shipped in a 4-week sprint, and shipped streaming Q&A with inline source cards, a usage-based paywall, and A/B-tested conversation variants in the React Native app.
Jun 2024 - Sep 2025
Turing
Python Developer (LLM Training)
LLM-training contractor; SFT training data for data-analysis tasks.
- Promoted to pod lead, guiding a 4-person pod and reporting on performance.
- Wrote scripts that tracked each pod member's contributions and generated the weekly progress reports for stakeholders.
May 2021 - Oct 2024
Keepoala
Software Developer
Core mobile engineer on a two-person mobile team, from working student onward.
- Built and shipped the React Native consumer app (iOS + Android) over 3.5 years, including a full UI/UX redesign pass and reusable design-system components.
- Backend integrations on Firebase Cloud Functions (BillBee, WeClapp, Algolia) and a Shopify merchant plugin.
- Partner-shop analytics dashboard (React + R-Shiny): retention curves, NPS, and order metrics.
Oct 2021 - Mar 2022; Jan 2025 - Apr 2025
Karat
Interview Engineer
Two contract stints as a technical interviewer.
- Interviewed software-engineering candidates on behalf of Karat's enterprise clients using calibrated rubrics: structured coding and problem-solving screens.
2017 - 2022
B.Sc. Computer Engineering · Cairo University
Very Good with honors · Graduation project: NeuroGames (graded Excellent)
Things I've built
What I work with
Solid = daily drivers
Mobile
LLMs / Fine-tuning
RAG / Retrieval
Agents
Evaluation & observability
Backend & Infra
AI-assisted dev
Languages
Let's talk.
Looking for applied AI, forward-deployed, or React Native roles, remote (CET/EU) or with relocation. Email me and I’ll usually reply the same day.