Principal Data Scientist - Generative AI
About the job
- 10–15+ years of progressive experience in Data Science, Machine Learning, and Applied Research, including 4+ years leading Generative AI and Deep Learning initiatives.
- Deep theoretical and practical mastery of LLM architectures, attention mechanisms, embedding spaces, fine-tuning techniques including PEFT, LoRA, and QLoRA, and model alignment including DPO and RLHF.
- A proven track record designing enterprise-grade RAG architectures, vector search ecosystems using Pinecone, Qdrant, OpenSearch, and pgvector, and hybrid retrieval algorithms.
- Industry expertise designing automated LLM evaluation frameworks such as RAGAS, DeepEval, and TruLens, and ground-truth dataset curation methods.
- Deep fluency in Python and machine-learning and deep-learning frameworks including PyTorch, Scikit-learn, Pandas, and NumPy.
- Experience deploying and optimizing models across Azure AI Foundry, AWS Bedrock, and Databricks/MLflow.
- A strong commitment to statistical validation, empirical testing, and evidence-based decision-making.
- The ability to connect technical metrics such as Normalized Discounted Cumulative Gain, BLEU/ROUGE, and context precision to return on investment and business key performance indicators.
- Clear and persuasive communication capable of building consensus among research scientists, platform engineers, and business leaders.
- Define the enterprise science roadmap for Generative AI, model alignment, advanced retrieval, and AI safety and trustworthiness.
- Pioneer RAG paradigms including GraphRAG, Knowledge Graphs, multimodal embeddings, and hybrid sparse/dense search architectures.
- Lead model selection, adaptation, and optimization strategies across proprietary and open-source foundation models, evaluating when to use RAG, fine-tuning through SFT/LoRA, or In-Context Learning.
- Partner with business executives and product leaders to translate business objectives into quantifiable AI metrics and scientific experiments.
- Establish enterprise-wide evaluation standards, benchmarking frameworks, and automated testbeds for hallucination detection, factual precision, bias mitigation, and toxicity prevention.
- Architect continuous monitoring and statistical evaluation mechanisms for live production systems, including online LLM observability, output drift monitoring, and semantic regression testing.
- Lead Responsible AI, Explainability, and AI Governance initiatives by defining scientific methodologies to audit AI decisions in highly regulated enterprise environments.
- Work with Enterprise AI Architects and AI Engineering teams to transition research prototypes into scalable, production-ready microservices.
- Mentor and upskill senior data scientists, applied machine-learning engineers, and offshore team members on advanced machine-learning methods, statistical evaluation rigor, and research best practices.
- Represent the enterprise as a subject-matter expert in AI innovation and present key technical breakthroughs to executive stakeholders.
- Hands-on expertise with GraphRAG, Knowledge Graphs, and graph databases such as Neo4j.
- A strong background in Banking, Financial Services, or similarly regulated enterprise domains.
- Experience with distributed computing frameworks such as PySpark and Databricks for large-scale data processing and model evaluation.
- A publication record or open-source contributions in applied Natural Language Processing, Information Retrieval, or Generative AI.
About the company
Photon designs, builds, and runs custom software for large companies, combining an in-house creative agency with its own engineering teams. Its flagship client work includes rebuilding Samsung's US online store as a set of smaller linked services and creating Olive Garden's mobile app, which lets diners join the waitlist from their phones instead of waiting at the door. Clients hire Photon's teams to plan and code their apps, websites, and AI tools and to keep them running afterward, paying for the engineers and designers assigned to each project. Photon is privately held and based in London, with its US arm in Dallas and engineering centers in Chennai and Bengaluru.
Company Size
5,001-10,000
Company Stage
N/A
Total Funding
N/A
Headquarters
London, United Kingdom
Founded
N/A
Get referred to Photon
See people who can refer or advise you
Simplify's Take
What believers are saying
- Photon is hiring at scale: it listed 940 open roles in October 2026, 551 of them posted since January.
- Headcount keeps climbing: Photon's engineers grew from over 5,000 in 2022 to more than 8,500 by August 2026.
- Visa sponsorship is routine: Photon made 194 H-1B labor filings, a work visa's first step, by October 8, 2026.
What critics are saying
- Developer pay runs below average: Photon's visa filings from January to October 2026 offered a median $110,750, versus $135,980 nationally.
- Most openings are offshore: Photon listed 613 open roles in India in October 2026, versus 291 in the US.
- Don't expect startup-style equity: Photon is privately held and grew on its own profits, without outside investors.
What makes Photon unique
- Its client list is unusually blue-chip: Photon says 40% of the Fortune 100 use it, including Walgreens and Wells Fargo.
- Security rigor is a selling point: an independent audit scored Photon's cybersecurity 987 of 1,000 in August 2026.
- Its work runs at consumer scale: Photon says apps it supports handle over 1 billion interactions a day.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Health Insurance
Dental Insurance
Vision Insurance
401(k) Retirement Plan
Paid Vacation
Paid Holidays
Performance Bonus
Company News
Photon raises millions to replace mobile apps amid SB 1047 controversy. The market is navigating a transition from static software interfaces to agentic systems. Photon raised $4.5M to develop an agent-based alternative to mobile apps, signaling a strategic bet on the obsolescence of traditional UI. This movement is supported by new research into autonomous ML... Executive summary. The market is navigating a transition from static software interfaces to agentic systems. Photon raised $4.5M to develop an agent-based alternative to mobile apps, signaling a strategic bet on the obsolescence of traditional UI. This movement is supported by new research into autonomous ML engineering, which indicates that labs are increasingly focused on systems that can manage their own development cycles. Infrastructure efficiency continues to improve as the Allen Institute for AI released Olmo-core 3 for scalable MoE training. This move lowers the cost of entry for specialized model development and complicates the competitive advantage for proprietary labs. Boards should track the widening gap between technical progress and safety definitions, as the current lack of a unified governance framework creates persistent regulatory risk for enterprise adoption.
Photon achieves CyberVadis Platinum Rating, reinforcing enterprise trust through independent cybersecurity validation. Sep 01, 2026, 10:00 ET Independent assessment score of 987 out of 1,000 validates the maturity of Photon's cybersecurity governance, operational resilience, and enterprise security program PARIS, Sept. 1, 2026 /PRNewswire/ - As enterprises accelerate investments in artificial intelligence, cloud modernization, and digital transformation, cybersecurity has become a defining measure of enterprise trust. Organizations increasingly expect technology partners to demonstrate mature security governance supported by independent validation - not simply technical capability. Against this backdrop, Photon today announced that it has achieved the CyberVadis Platinum Rating, receiving an exceptional 987 out of 1,000 in the 2026 CyberVadis cybersecurity assessment. The Platinum Rating - the highest recognition level within the CyberVadis assessment framework - provides independent, evidence-based validation of the organization's cybersecurity maturity, recognizing the strength of its governance, operational controls, risk management practices, and continuous improvement program. For organizations across banking, financial services, insurance, healthcare, retail, manufacturing, and other highly regulated industries, cybersecurity is no longer simply about protecting technology. It is fundamental to business resilience, responsible AI adoption, regulatory confidence, and sustainable digital transformation. "Trust is the foundation of every transformation we deliver," said Srinivas Balasubramanian, Co-Founder, Chairman and Chief Executive Officer of Photon. "Our clients rely on Photon to modernize mission-critical platforms, engineer AI-powered experiences, and operate secure digital ecosystems at enterprise scale. Achieving the CyberVadis Platinum Rating with a score of 987 provides independent validation of the governance, operational discipline, and security culture that underpin every engagement. As AI reshapes the enterprise, we will continue investing in the cybersecurity capabilities that enable our clients to innovate with confidence." Photon's 2026 result builds on an already exceptional prior assessment, when the company achieved a CyberVadis score of 953 out of 1,000, placing it among the top 1% of companies worldwide. Improving to 987 out of 1,000 makes this year's result especially significant, demonstrating year-over-year progress from an already industry-leading position and reinforcing Photon's standing among the top tier of global cybersecurity maturity. Independent Validation of Enterprise Cybersecurity Maturity CyberVadis provides evidence-based third-party cybersecurity assessments supported by independent cybersecurity analyst review. Rather than evaluating individual technologies or products, its methodology measures how effectively organizations establish, implement, monitor, and continuously improve enterprise cybersecurity programs through documented governance, operational controls, technical safeguards, and measurable execution. The assessment evaluates organizations across four foundational cybersecurity functions: * Identify: Governance, asset management, cybersecurity strategy, enterprise risk, regulatory compliance, and third-party oversight. * Protect: Identity and access management, infrastructure security, data protection, employee awareness, and preventive controls. * Detect: Security monitoring, vulnerability management, threat detection, and continuous visibility. * React: Incident response, recovery planning, crisis management, business continuity, and continuous improvement. Photon's assessment score of 987 out of 1,000 reflects a highly mature cybersecurity operating environment supported by disciplined governance, comprehensive security controls, and consistently executed operational practices. Security Embedded Into Every Transformation At Photon, security is not a standalone capability - it is embedded into every transformation program the company delivers. Across AI, cloud, digital commerce, customer experience, software engineering, and enterprise modernization, Photon integrates cybersecurity into strategy, architecture, engineering, deployment, and operations, enabling organizations to accelerate innovation while maintaining the governance, resilience, and operational control required for enterprise-scale transformation. Photon's enterprise cybersecurity capabilities include: * Application Security * Identity and Access Management * Governance, Risk and Compliance * Cloud Security * Threat Management * Endpoint Security * Security Operations and Monitoring * Infrastructure Security * Incident Response * Business Continuity and Operational Resilience For enterprise clients, the CyberVadis assessment provides an independent benchmark that supports supplier due diligence, third-party risk management, procurement evaluations, and governance discussions while complementing each organization's own cybersecurity assessment processes. Building the Trusted Foundation for AI Artificial intelligence is fundamentally changing enterprise risk. As organizations deploy intelligent agents, automate decision-making, and expand cloud-native operations, cybersecurity must extend beyond perimeter protection to encompass governance, identity, data protection, observability, model integrity, and operational accountability. Photon believes trust is the foundation of every transformation it delivers. By embedding security, governance, and operational resilience into every engagement, Photon helps organizations accelerate AI adoption while maintaining the transparency, control, and confidence required for enterprise-scale deployment. The CyberVadis Platinum Rating reinforces Photon's ongoing commitment to continuously strengthening the security capabilities, governance practices, and operational excellence that enable clients to transform with confidence. About CyberVadis CyberVadis provides evidence-based third-party cybersecurity assessments through a cloud-based platform supported by independent cybersecurity analyst review. Organizations worldwide use CyberVadis assessments to evaluate supplier cybersecurity maturity, strengthen third-party risk management, and improve transparency across global digital ecosystems. Its methodology produces an overall cybersecurity maturity score ranging from 0 to 1,000 together with detailed evaluations across the four core cybersecurity functions of Identify, Protect, Detect, and React. About Photon Photon is an outcome-driven AI and digital transformation company with more than 8,500 engineers serving over 40% of the Fortune 100 across Banking, Financial Services, Retail, Consumer Products, Healthcare, Automotive, and other industries. For more than 25 years, Photon has partnered with the world's leading enterprises to modernize customer experiences, engineer intelligent platforms, and accelerate business transformation through AI, data, cloud, software engineering, cybersecurity, and digital operations. Combining strategic advisory, deep engineering expertise, and enterprise-scale delivery, Photon helps organizations move from AI ambition to production with speed, security, and measurable business outcomes. With delivery centers across North America, Europe, and Asia, Photon delivers the governance, engineering discipline, and operational excellence required to build the next generation of intelligent enterprises. SOURCE Photon
Exein unveils next-generation runtime security to protect the ai-native world. * Photon blocks cyberattacks before execution across physical AI and IoT, autonomous AI agents and cloud and edge infrastructure * Kernel-level prevention sets a new standard beyond traditional user-space detection * Builds on Exein's position as the world's largest runtime security provider, protecting over two billion devices Mar 23, 2026 Prev Next 1 of 42,872 Exein, the global leader in runtime cybersecurity, unveiled Photon, a preemptive breakthrough solution that blocks cyberattacks at the point of execution. Designed for the AI-native world - where digital and physical systems are now inseparable - Photon marks a fundamental shift in how critical infrastructure protects itself. Unlike traditional cybersecurity solutions that detect threats after compromise - typically operating in user space and relying on a cloud network - Exein's Photon operates directly inside the kernel, preventing malicious execution paths before they can run. By blocking attacks before the point of execution, the technology dramatically reduces latency and eliminates entire classes of threats before damage occurs. If malicious instructions cannot execute, the attack itself cannot take place. This advancement establishes a new category of runtime security designed for systems that cannot be disconnected: physical AI and IoT environments, autonomous AI agents, and local hybrid cloud and edge infrastructure. In these environments, from industrial robotics and critical infrastructure to AI-driven platforms, downtime is not an option, and protection must be more precise and granular, blocking malicious threats without shutting down the entire process. The announcement at the RSA Conference (RSAC) comes as cyber threats increasingly target physical systems. Last month, the Munich Security Report 2026 warned that cyber operations are now engineered to cause real-world disruption, accelerating regulatory intervention after voluntary measures failed to address systemic vulnerabilities. At the same time, the speed of attacks is accelerating dramatically: recent threat intelligence shows average attacker 'breakout times' fell to just 29 minutes in 2025, 65% faster than the previous year, driven in part by AI-assisted automation. Protecting the digital and physical in the AI era Artificial intelligence is already capable of identifying vulnerabilities in software and infrastructure. In the near future, these models will not only detect weaknesses but exploit them autonomously to launch attacks at machine speed. As the scale and sophistication of these attacks grow, traditional runtime security systems that rely on detection alone will no longer be sufficient. Photon introduces a new model of preemptive runtime security designed for this AI-driven environment. Rather than detecting attacks after they begin, it prevents malicious execution paths from running in the first place, blocking threats in real time before they can impact the system. Unlike conventional security tools that operate in user space alongside the applications they protect, Photon operates directly within the kernel, the core of the operating system. By enforcing protection at this foundational layer, rather than merely detecting and stopping attacks, it prevents them from executing in the first place - all in real time. This marks a major milestone as physical and digital systems converge, positioning Photon as a new reference architecture for securing physical AI, agent AI and cloud and hybrid infrastructure. Gianni Cuozzo, Founder and CEO of Exein, said: "In a future where the world is infinitely connected with humanoid robots walking among Aithority, local LLMs powering intelligent edges, autonomous drones reshaping mobility, and billions of new autonomous systems bridging the digital and physical realms, preemptive runtime security represents the new generation of protection, built into the very DNA of every device from the ground up. "Exein was born to make this vision a reality: transforming every connected device into a fortress of security, forging the largest decentralised immune system for digital life - cross-vendor, cross-platform, and cross-system. We stand as the first line of defence between the boundless digital world and the physical one we live in, empowering manufacturers to build inherently safe innovations and already safeguarding over 2 billion devices worldwide."